diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 2fd24a7e4836162074b7646d91d30a11a1c9dd27..76277ffafcaa1a34d5838e0ef68ebae21626218c 100644 --- a/analysis/current/analysis-report-hybrid.json +++ b/analysis/current/analysis-report-hybrid.json @@ -1,11 +1,567 @@ { "schema_version": "1.0", "repo": "huggingface/transformers", - "snapshot_id": "20260629T180016Z", - "generated_at": "2026-06-29T18:11:39Z", + "snapshot_id": "20260630T000016Z", + "generated_at": "2026-06-30T00:09:40Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ + { + "cluster_id": "cluster-43451-8", + "summary": "Cluster of 8 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 43451, + "issue_numbers": [], + "pr_numbers": [ + 43451, + 46658, + 46707, + 46836, + 46850, + 46891, + 46953, + 46957 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43451, + "right_pr_number": 46658, + "code_similarity": 0.243, + "size_similarity": 0.955, + "file_overlap": 0.027, + "area_overlap": 0.0, + "patch_similarity": 0.293, + "shared_filenames": [ + "src/transformers/models/auto/auto_mappings.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 385, + 392 + ], + [ + 813, + 819 + ], + [ + 933, + 939 + ], + [ + 1054, + 1060 + ], + [ + 1186, + 1192 + ] + ], + "right_ranges": [ + [ + 362, + 368 + ], + [ + 591, + 597 + ], + [ + 796, + 802 + ] + ] + } + ] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46707, + "code_similarity": 0.026, + "size_similarity": 0.071, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.076, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46836, + "code_similarity": 0.238, + "size_similarity": 0.715, + "file_overlap": 0.152, + "area_overlap": 0.0, + "patch_similarity": 0.33, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "src/transformers/conversion_mapping.py", + "src/transformers/models/__init__.py", + "src/transformers/models/auto/auto_mappings.py", + "src/transformers/models/auto/modeling_auto.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 1361, + 1368 + ] + ], + "right_ranges": [ + [ + 885, + 892 + ] + ] + }, + { + "filename": "src/transformers/conversion_mapping.py", + "left_ranges": [ + [ + 627, + 659 + ] + ], + "right_ranges": [ + [ + 1380, + 1446 + ] + ] + }, + { + "filename": "src/transformers/models/__init__.py", + "left_ranges": [ + [ + 288, + 294 + ] + ], + "right_ranges": [ + [ + 450, + 456 + ] + ] + }, + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 385, + 392 + ], + [ + 813, + 819 + ], + [ + 933, + 939 + ], + [ + 1054, + 1060 + ], + [ + 1186, + 1192 + ] + ], + "right_ranges": [ + [ + 623, + 633 + ], + [ + 898, + 907 + ], + [ + 1102, + 1108 + ], + [ + 1228, + 1234 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 325, + 332 + ], + [ + 1081, + 1087 + ], + [ + 1900, + 1906 + ] + ], + "right_ranges": [ + [ + 495, + 501 + ], + [ + 1097, + 1103 + ] + ] + } + ] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46850, + "code_similarity": 0.027, + "size_similarity": 0.093, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.055, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46891, + "code_similarity": 0.002, + "size_similarity": 0.007, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.007, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46953, + "code_similarity": 0.002, + "size_similarity": 0.01, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.003, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46957, + "code_similarity": 0.012, + "size_similarity": 0.029, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.042, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46707, + "code_similarity": 0.027, + "size_similarity": 0.074, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.079, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46836, + "code_similarity": 0.205, + "size_similarity": 0.748, + "file_overlap": 0.029, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [ + "src/transformers/models/auto/auto_mappings.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 362, + 368 + ], + [ + 591, + 597 + ], + [ + 796, + 802 + ] + ], + "right_ranges": [ + [ + 623, + 633 + ], + [ + 898, + 907 + ], + [ + 1102, + 1108 + ], + [ + 1228, + 1234 + ] + ] + } + ] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46850, + "code_similarity": 0.028, + "size_similarity": 0.098, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.054, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46891, + "code_similarity": 0.003, + "size_similarity": 0.007, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.008, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46953, + "code_similarity": 0.003, + "size_similarity": 0.011, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.003, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46957, + "code_similarity": 0.013, + "size_similarity": 0.03, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.046, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46836, + "code_similarity": 0.031, + "size_similarity": 0.099, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.078, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46850, + "code_similarity": 0.161, + "size_similarity": 0.76, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.058, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46891, + "code_similarity": 0.021, + "size_similarity": 0.093, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.014, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46953, + "code_similarity": 0.031, + "size_similarity": 0.143, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.015, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46957, + "code_similarity": 0.121, + "size_similarity": 0.403, + "file_overlap": 0.1, + "area_overlap": 0.0, + "patch_similarity": 0.069, + "shared_filenames": [ + "src/transformers/generation/utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/generation/utils.py", + "left_ranges": [ + [ + 33, + 46 + ] + ], + "right_ranges": [ + [ + 2156, + 2186 + ], + [ + 2865, + 2871 + ] + ] + } + ] + }, + { + "left_pr_number": 46836, + "right_pr_number": 46850, + "code_similarity": 0.035, + "size_similarity": 0.131, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.057, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46836, + "right_pr_number": 46891, + "code_similarity": 0.003, + "size_similarity": 0.009, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.008, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46836, + "right_pr_number": 46953, + "code_similarity": 0.003, + "size_similarity": 0.014, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.003, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46836, + "right_pr_number": 46957, + "code_similarity": 0.015, + "size_similarity": 0.04, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.049, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46850, + "right_pr_number": 46891, + "code_similarity": 0.017, + "size_similarity": 0.071, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.019, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46850, + "right_pr_number": 46953, + "code_similarity": 0.027, + "size_similarity": 0.109, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.033, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46850, + "right_pr_number": 46957, + "code_similarity": 0.071, + "size_similarity": 0.306, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.068, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46891, + "right_pr_number": 46953, + "code_similarity": 0.13, + "size_similarity": 0.652, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.0, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46891, + "right_pr_number": 46957, + "code_similarity": 0.049, + "size_similarity": 0.231, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.019, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46953, + "right_pr_number": 46957, + "code_similarity": 0.075, + "size_similarity": 0.355, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.027, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, { "cluster_id": "cluster-43979-11", "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", @@ -988,371 +1544,62 @@ ], [ 658, - 663 - ], - [ - 666, - 674 - ], - [ - 717, - 722 - ], - [ - 738, - 744 - ], - [ - 749, - 772 - ], - [ - 790, - 795 - ] - ] - } - ] - }, - { - "left_pr_number": 44129, - "right_pr_number": 44154, - "code_similarity": 0.099, - "size_similarity": 0.344, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.199, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 44129, - "right_pr_number": 44722, - "code_similarity": 0.146, - "size_similarity": 0.545, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.247, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 44154, - "right_pr_number": 44722, - "code_similarity": 0.164, - "size_similarity": 0.631, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.255, - "shared_filenames": [], - "shared_file_areas": [] - } - ] - }, - { - "cluster_id": "cluster-43451-6", - "summary": "Cluster of 6 related pull requests linked by soft_similarity.", - "status": "open", - "confidence": 0.5, - "canonical_issue_number": null, - "canonical_pr_number": 43451, - "issue_numbers": [], - "pr_numbers": [ - 43451, - 46836, - 46891, - 46953, - 46957, - 46961 - ], - "evidence_types": [ - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 43451, - "right_pr_number": 46836, - "code_similarity": 0.238, - "size_similarity": 0.715, - "file_overlap": 0.152, - "area_overlap": 0.0, - "patch_similarity": 0.33, - "shared_filenames": [ - "docs/source/en/_toctree.yml", - "src/transformers/conversion_mapping.py", - "src/transformers/models/__init__.py", - "src/transformers/models/auto/auto_mappings.py", - "src/transformers/models/auto/modeling_auto.py" - ], - "shared_file_areas": [ - { - "filename": "docs/source/en/_toctree.yml", - "left_ranges": [ - [ - 1361, - 1368 - ] - ], - "right_ranges": [ - [ - 885, - 892 - ] - ] - }, - { - "filename": "src/transformers/conversion_mapping.py", - "left_ranges": [ - [ - 627, - 659 - ] - ], - "right_ranges": [ - [ - 1380, - 1446 - ] - ] - }, - { - "filename": "src/transformers/models/__init__.py", - "left_ranges": [ - [ - 288, - 294 - ] - ], - "right_ranges": [ - [ - 450, - 456 - ] - ] - }, - { - "filename": "src/transformers/models/auto/auto_mappings.py", - "left_ranges": [ - [ - 385, - 392 - ], - [ - 813, - 819 - ], - [ - 933, - 939 - ], - [ - 1054, - 1060 - ], - [ - 1186, - 1192 - ] - ], - "right_ranges": [ - [ - 623, - 633 - ], - [ - 898, - 907 - ], - [ - 1102, - 1108 + 663 ], [ - 1228, - 1234 - ] - ] - }, - { - "filename": "src/transformers/models/auto/modeling_auto.py", - "left_ranges": [ - [ - 325, - 332 + 666, + 674 ], [ - 1081, - 1087 + 717, + 722 ], [ - 1900, - 1906 - ] - ], - "right_ranges": [ + 738, + 744 + ], [ - 495, - 501 + 749, + 772 ], [ - 1097, - 1103 + 790, + 795 ] ] } ] }, { - "left_pr_number": 43451, - "right_pr_number": 46891, - "code_similarity": 0.002, - "size_similarity": 0.007, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.007, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 43451, - "right_pr_number": 46953, - "code_similarity": 0.002, - "size_similarity": 0.01, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.003, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 43451, - "right_pr_number": 46957, - "code_similarity": 0.012, - "size_similarity": 0.029, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.042, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 43451, - "right_pr_number": 46961, - "code_similarity": 0.004, - "size_similarity": 0.007, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.019, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46836, - "right_pr_number": 46891, - "code_similarity": 0.003, - "size_similarity": 0.009, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.008, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46836, - "right_pr_number": 46953, - "code_similarity": 0.003, - "size_similarity": 0.014, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.003, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46836, - "right_pr_number": 46957, - "code_similarity": 0.015, - "size_similarity": 0.04, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.049, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46836, - "right_pr_number": 46961, - "code_similarity": 0.005, - "size_similarity": 0.01, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.019, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46891, - "right_pr_number": 46953, - "code_similarity": 0.13, - "size_similarity": 0.652, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.0, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46891, - "right_pr_number": 46957, - "code_similarity": 0.049, - "size_similarity": 0.231, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.019, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46891, - "right_pr_number": 46961, - "code_similarity": 0.202, - "size_similarity": 0.935, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.102, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 46953, - "right_pr_number": 46957, - "code_similarity": 0.075, - "size_similarity": 0.355, + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.027, + "patch_similarity": 0.199, "shared_filenames": [], "shared_file_areas": [] }, { - "left_pr_number": 46953, - "right_pr_number": 46961, - "code_similarity": 0.141, - "size_similarity": 0.697, + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.01, + "patch_similarity": 0.247, "shared_filenames": [], "shared_file_areas": [] }, { - "left_pr_number": 46957, - "right_pr_number": 46961, - "code_similarity": 0.053, - "size_similarity": 0.247, + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, "file_overlap": 0.0, "area_overlap": 0.0, - "patch_similarity": 0.022, + "patch_similarity": 0.255, "shared_filenames": [], "shared_file_areas": [] } @@ -1458,20 +1705,18 @@ ] }, { - "cluster_id": "cluster-46894-6", - "summary": "Cluster of 3 issues and 3 PRs centered on issue #46894.", + "cluster_id": "cluster-46894-4", + "summary": "Cluster of 2 issues and 2 PRs centered on issue #46894.", "status": "open", "confidence": 0.8, "canonical_issue_number": 46894, - "canonical_pr_number": 46903, + "canonical_pr_number": 46895, "issue_numbers": [ 46894, - 46897, - 46901 + 46897 ], "pr_numbers": [ 46895, - 46903, 46958 ], "evidence_types": [ @@ -1480,59 +1725,6 @@ "soft_similarity" ], "pr_comparisons": [ - { - "left_pr_number": 46895, - "right_pr_number": 46903, - "code_similarity": 0.362, - "size_similarity": 0.443, - "file_overlap": 0.333, - "area_overlap": 0.419, - "patch_similarity": 0.178, - "shared_filenames": [ - "src/transformers/models/cohere_asr/modeling_cohere_asr.py", - "src/transformers/models/cohere_asr/modular_cohere_asr.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", - "left_ranges": [ - [ - 615, - 627 - ], - [ - 639, - 654 - ] - ], - "right_ranges": [ - [ - 637, - 651 - ] - ] - }, - { - "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", - "left_ranges": [ - [ - 478, - 490 - ], - [ - 502, - 517 - ] - ], - "right_ranges": [ - [ - 500, - 514 - ] - ] - } - ] - }, { "left_pr_number": 46895, "right_pr_number": 46958, @@ -1617,59 +1809,6 @@ ] } ] - }, - { - "left_pr_number": 46903, - "right_pr_number": 46958, - "code_similarity": 0.325, - "size_similarity": 0.369, - "file_overlap": 0.333, - "area_overlap": 0.364, - "patch_similarity": 0.156, - "shared_filenames": [ - "src/transformers/models/cohere_asr/modeling_cohere_asr.py", - "src/transformers/models/cohere_asr/modular_cohere_asr.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", - "left_ranges": [ - [ - 637, - 651 - ] - ], - "right_ranges": [ - [ - 22, - 28 - ], - [ - 638, - 645 - ] - ] - }, - { - "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", - "left_ranges": [ - [ - 500, - 514 - ] - ], - "right_ranges": [ - [ - 16, - 22 - ], - [ - 501, - 508 - ] - ] - } - ] } ] }, @@ -2701,13 +2840,12 @@ ], "duplicate_issues": [ { - "cluster_id": "cluster-46894-6", + "cluster_id": "cluster-46894-4", "canonical_issue_number": 46894, "duplicate_issue_numbers": [ - 46897, - 46901 + 46897 ], - "reason": "Issues in cluster-46894-6 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." + "reason": "Issues in cluster-46894-4 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." } ], "duplicate_prs": [ @@ -2730,17 +2868,19 @@ "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." }, { - "cluster_id": "cluster-43451-6", + "cluster_id": "cluster-43451-8", "canonical_pr_number": 43451, "duplicate_pr_numbers": [ + 46658, + 46707, 46836, + 46850, 46891, 46953, - 46957, - 46961 + 46957 ], "target_issue_number": null, - "reason": "PRs in cluster-43451-6 are treated as duplicates because they share soft_similarity evidence." + "reason": "PRs in cluster-43451-8 are treated as duplicates because they share soft_similarity evidence." }, { "cluster_id": "cluster-43698-3", @@ -2832,21 +2972,20 @@ "reason": "PRs in cluster-46869-3 are treated as duplicates because they converge on issue #46869 with closing_reference, shared_issue_target evidence." }, { - "cluster_id": "cluster-46894-6", - "canonical_pr_number": 46903, + "cluster_id": "cluster-46894-4", + "canonical_pr_number": 46895, "duplicate_pr_numbers": [ - 46895, 46958 ], "target_issue_number": 46894, - "reason": "PRs in cluster-46894-6 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." + "reason": "PRs in cluster-46894-4 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." } ], "best_issue": { - "cluster_id": "cluster-46894-6", + "cluster_id": "cluster-46894-4", "issue_number": 46894, - "reason": "Issue #46894 is the strongest global issue candidate because it is open, belongs to a cluster with 6 artifacts, and carries 0 discussion comments plus 6 inbound references.", - "score": 166.0 + "reason": "Issue #46894 is the strongest global issue candidate because it is open, belongs to a cluster with 4 artifacts, and carries 0 discussion comments plus 6 inbound references.", + "score": 146.0 }, "best_pr": { "cluster_id": "cluster-43979-11", diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index d597be0df306c9eee1eca0818954bd69c3e5e8b0..118778e3a235272cc0aa35d9d0e2b115c7b62c32 100644 --- a/analysis/current/analysis-report-hybrid.llm-reviews.json +++ b/analysis/current/analysis-report-hybrid.llm-reviews.json @@ -1,8 +1,8 @@ { "schema_version": "1.0", "repo": "huggingface/transformers", - "snapshot_id": "20260629T180016Z", - "generated_at": "2026-06-29T18:11:39Z", + "snapshot_id": "20260630T000016Z", + "generated_at": "2026-06-30T00:09:40Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -132,412 +132,382 @@ "issue:45237|issue:46015", "issue:43039|issue:44805" ], - "prepared_review_unit_hash": "159f3ce72dc6e372186ca9467d0231907ef6fb07f777a634640bb1a29aefa19f", + "prepared_review_unit_hash": "47580346ddada2b540346b57f447a3c240eb25310bae3ce3ca62120f5963ddc6", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 63, - "serialized_chars": 46573, - "estimated_input_tokens": 11644, - "estimated_eval_tokens": 23544 + "serialized_chars": 46571, + "estimated_input_tokens": 11643, + "estimated_eval_tokens": 23542 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly noisy soft-similarity matches across unrelated issues. The only strong duplicate-looking pair is the ZeRO-3/BERT index-out-of-bounds reports, with a second plausible torch-compatibility pair.", - "confidence": 0.87, - "canonical_issue_reason": "Issue 43596 is the clearest concrete bug report in the set and has the closest true duplicate (43638); both describe the same ZeRO-3/BERT index-out-of-bounds failure during model loading.", + "summary": "This cluster is mostly heterogeneous: a few true duplicates around ZeRO-3 / torch-version / Qwen3.5-MoE rotary-embedding issues, but many other pairs are only loosely related model/runtime failures or docs regressions and should stay separate.", + "confidence": 0.56, + "canonical_issue_reason": "Issue 45237 is the best cluster representative because it sits in the densest subgroup of model/device-placement failures (GPU, auto device map, offload, MoE/runtime errors) and is broad enough to summarize the dominant theme.", "canonical_pr_reason": null, - "best_issue_reason": "43596 is the best representative issue because it is specific, reproducible, and directly tied to a near-identical follow-up report.", + "best_issue_reason": "45237 is the most suitable issue to keep as the global representative: it is concrete, central to the cluster\u2019s main runtime/device-mapping theme, and broader than the more model-specific outliers.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different model/API failures: Qwen3.5 config attribute error vs streaming chat/completions failure." + "reason": "Different failures: Qwen3.5 config attribute vs streaming chat/completions API." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different runtime paths: AMD GPU device-map failure vs MPS histogram op not implemented." + "reason": "Different concrete bugs: AMD GPU failure vs MPS histogram kernel limitation." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different models and bugs: Mistral dtype mismatch vs GPT-OSS on AMD GPUs." + "reason": "Both are accelerator-related, but they affect different model paths and error types." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both are runtime errors, but the model, device context, and failure mode differ." + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is a different code path from GPT-OSS AMD GPU failure." }, { "left": "issue:45237", "right": "issue:46468", "accept": false, - "reason": "Unrelated model/runtime issue vs broken Nemotron example." + "reason": "Different models and symptoms; no shared concrete bug." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different problems: GPU device-map failure vs invalid method call in GraniteMoEHybrid." + "reason": "GPU/runtime failure vs invalid method call in GraniteMoEHybrid; not the same bug." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch and ONNX export error are unrelated." + "reason": "DeepSpeed collective mismatch is unrelated to ONNX export." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "ModernBERT ONNX export and Qwen3Moe FP8 failures are different bugs." + "reason": "ModernBERT ONNX export and Qwen3Moe FineGrainedFP8 are unrelated." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "Different subsystems: export-time ONNX issue vs Qwen3 auto device-map assert." + "reason": "ONNX export error vs Qwen3 auto device-map assert are different problems." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Table QA crash and RAG example docs are unrelated." + "reason": "Table QA crash and RAG docs cleanup are unrelated." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs cleanup items for v5, but they target different obsolete content." + "reason": "Both are docs regressions, but they reference different removed v5 features and are not the same issue." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "Different failure classes: ONNX export vs missing GenerationMixin import." + "reason": "ONNX export failure is unrelated to GenerationMixin import breakage." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "Different training issues: TRL SFT embedding unpacking vs resume/trust-remote-code behavior." + "reason": "TRL SFT embedding unpacking vs resume-training with remote code are different failures." }, { "left": "issue:46165", "right": "issue:46869", "accept": false, - "reason": "Resume-training problem and YOLO multi-GPU device-map failure are unrelated." + "reason": "Resume-training issue is unrelated to YOLO multi-GPU device-map failure." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Mask-shape IndexError and Gemma4 PLE device mismatch are different bugs." + "reason": "Mask/tensor shape mismatch is not the same as Gemma4 PLE device mismatch." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Different error signatures and code paths." + "reason": "Both are IndexErrors, but the tensor shape mismatch is a different failure mode from the ZeRO-3 empty-index bug." }, { - "left": "issue:43638", + "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "ZeRO-3 BERT loading bug vs resume-training issue; not the same bug." + "reason": "ZeRO-3 Bert index error is unrelated to resume-training with local models." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Both are IndexErrors, but the model/loading path and root symptom differ." + "reason": "Similar IndexError wording, but the underlying bugs are different." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask/tensor shape mismatch is unrelated to Qwen3.5 video template mismatch." + "reason": "Mask shape mismatch is unrelated to Qwen3.5 video token/feature mismatch." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Different model/runtime problems: mask shape error vs GPT-OSS TP issue." + "reason": "Different models and different device/matrix errors." }, { "left": "issue:46165", "right": "issue:46468", "accept": false, - "reason": "Resume-training issue and Nemotron example breakage are unrelated." + "reason": "Training resume problem is unrelated to Nemotron example breakage." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video StopIteration bug and resume-training bug are unrelated." + "reason": "Video input StopIteration and training resume are unrelated." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice preset and ModernBERT ONNX export are unrelated." + "reason": "BarkProcessor voice_preset bug is unrelated to ONNX export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both involve DeepSpeed, but the concrete failures and affected models differ." + "reason": "Both involve DeepSpeed, but they are different concrete failures and code paths." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same ZeRO-3/BERT index-out-of-bounds failure during model loading; nearly identical reports." + "reason": "Same ZeRO-3/Bert empty-index failure; titles differ only by model phrasing." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger cross-entropy call issue and GPT-OSS tensor-parallel issue are unrelated." + "reason": "Liger Kernel cross_entropy call and GPT-OSS TP failure are unrelated." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "DeepSeek detokenization regression and LoRA merge/save mismatch are different bugs." + "reason": "Different problems: detokenization bug vs LoRA merge/save mismatch." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen NVFP4 runtime error and LoRA save/load output mismatch are unrelated." + "reason": "Qwen3-8B NVFP4 runtime error is not the same as LoRA save/reload inconsistency." }, { "left": "issue:44841", "right": "issue:46661", "accept": false, - "reason": "Processor failure for Voxtral and remote-code config_class AttributeError are different issues." + "reason": "Processor failure for Voxtral is unrelated to remote-code model_class config_class handling." }, { "left": "issue:43716", "right": "issue:46869", "accept": false, - "reason": "Different model families and different device-placement failures." + "reason": "Different models and failure modes; only broad device issues overlap." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation-checkpointing/FSDP tensor-size mismatch is not the same as the mask index error." + "reason": "FSDP checkpointing recompute mismatch is not the same as the mask/tensor shape IndexError." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Both mention DeepSpeed, but one is a 2-D weight training error and the other is a collective mismatch." + "reason": "DeepSpeed weight-dimension error and collective mismatch are distinct bugs." }, { "left": "issue:45685", "right": "issue:46869", "accept": false, - "reason": "MPS int histogram limitation and YOLO multi-GPU device-map failure are unrelated." + "reason": "MPS histogram issue is unrelated to YOLO multi-GPU device-map auto." }, { "left": "issue:45859", "right": "issue:45902", - "accept": false, - "reason": "Both touch Qwen3.5 MoE rotary embedding, but they describe different concrete bugs and triggers." + "accept": true, + "reason": "Same Qwen3_5MoeTextRotaryEmbedding offload/materialization bug surface; one report emphasizes CPU offload, the other uninitialized inv_freq state." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "BERT ZeRO-3 loading error and GPT-OSS tensor-parallel issue are unrelated." + "reason": "Both are model/runtime issues, but Bert ZeRO-3 index error and GPT-OSS TP failure are not the same bug." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "FSDP checkpoint tensor-size mismatch and CustomTrainer signature mismatch are different bugs." + "reason": "Activation-checkpointing/FSDP mismatch is unrelated to the CustomTrainer compute_loss signature change." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "ZeRO-3 BERT loading issue and GPT-OSS TP issue are unrelated." + "reason": "ZeRO-3 Bert index error is unrelated to GPT-OSS tensor-parallel failure." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Different models and failure modes: dtype mismatch vs invalid method call." + "reason": "Mistral-3 dtype mismatch and GraniteMoEHybrid invalid method are different issues." }, { "left": "issue:45874", "right": "issue:46294", "accept": false, - "reason": "Gemma4 reverse-embedding lookup and Qwen3.5 video token mismatch are different problems." + "reason": "Gemma4 reverse-embedding lookup bug is unrelated to Qwen3.5 video token mismatch." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "NVFP4 runtime failure and GPT-OSS AMD GPU issue are unrelated." + "reason": "Qwen3-8B NVFP4 runtime error and GPT-OSS AMD GPU failure are different model/runtime bugs." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Different model-specific regressions: Gemma3 config attribute error vs Mistral regex patch crash." + "reason": "Gemma3 sliding_window_pattern/vLLM regression is unrelated to Mistral regex patch crash." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Same model family, but device mismatch and expensive reverse embedding lookup are different bugs." + "reason": "Gemma4 PLE device mismatch and reverse embedding lookup are separate bugs." }, { "left": "issue:45800", "right": "issue:46515", "accept": true, - "reason": "Both report the same torch 2.4.1 compatibility regression; one is just phrased more generally." + "reason": "Same torch 2.4.1 incompatibility report, just phrased differently." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Generic torch incompatibility and Qwen NVFP4 runtime failure are not the same bug." + "reason": "Torch version incompatibility and Qwen3 NVFP4 runtime error are not the same issue." }, { "left": "issue:34928", "right": "issue:41720", "accept": false, - "reason": "FSDP activation-checkpointing mismatch and Qwen3 auto device-map assert are unrelated." - }, - { - "left": "issue:43828", - "right": "issue:45161", - "accept": false, - "reason": "Different code paths: autocast dtype mismatch vs GPT-OSS tensor-parallel failure." + "reason": "FSDP activation-checkpointing mismatch is unrelated to Qwen3 auto device-map assert." }, { "left": "issue:43828", - "right": "issue:44292", + "right": "issue:45685", "accept": false, - "reason": "Both are runtime errors, but they affect different models and failure mechanisms." + "reason": "Autocast dtype mismatch is unrelated to the MPS histogram kernel limitation." }, { "left": "issue:43541", "right": "issue:43716", "accept": false, - "reason": "Mixtral torch-dynamo grouped_mm failure and Mistral image-preprocessor dtype mismatch are unrelated." + "reason": "Mixtral grouped_mm torch-dynamo tracing error is unrelated to Mistral-3 dtype mismatch." }, { "left": "issue:43716", "right": "issue:46468", "accept": false, - "reason": "Different models and different failure surfaces." + "reason": "Different models and failure modes; not the same underlying bug." }, { "left": "issue:44292", "right": "issue:44918", "accept": false, - "reason": "Qwen NVFP4 runtime error and TRL SFT embedding-unpacking issue are unrelated." + "reason": "Qwen3 NVFP4 runtime error and TRL SFT embedding unpacking are distinct." }, { "left": "issue:45507", "right": "issue:45685", "accept": false, - "reason": "GraniteMoE invalid method and MPS int histogram limitation are unrelated." + "reason": "GraniteMoEHybrid invalid method call is unrelated to MPS Int histogram support." }, { "left": "issue:43716", "right": "issue:44292", "accept": false, - "reason": "Different model-specific runtime failures." + "reason": "Different models and errors; only broad runtime similarity." }, { "left": "issue:44918", "right": "issue:45161", "accept": false, - "reason": "TRL SFT embedding issue and GPT-OSS tensor-parallel issue are unrelated." + "reason": "TRL embedding unpacking failure is unrelated to GPT-OSS tensor-parallel failure." }, { "left": "issue:41762", "right": "issue:43638", "accept": false, - "reason": "Both are ZeRO-3 IndexErrors, but they affect different models and are not clearly the same bug." + "reason": "Both are ZeRO-3 IndexErrors, but they involve different models and likely different root causes." }, { "left": "issue:43039", "right": "issue:44918", "accept": false, - "reason": "Liger kernel cross-entropy issue and TRL SFT embedding issue are different bugs." + "reason": "Liger Kernel cross_entropy call and Qwen3.5 SFT trainer error are unrelated." }, { "left": "issue:43827", "right": "issue:44512", "accept": false, - "reason": "Both are docs issues, but they refer to different removed v5 content." + "reason": "Both are docs regressions, but they concern different removed v5 topics and should not be merged as one bug." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Autocast dtype mismatch and MPS integer histogram limitation are unrelated." - }, - { - "left": "issue:43039", - "right": "issue:43828", - "accept": false, - "reason": "Different model/runtime bugs with no shared concrete failure path." - }, - { - "left": "issue:30277", - "right": "issue:36331", - "accept": false, - "reason": "DeepSpeed collective mismatch and CustomTrainer argument mismatch are unrelated." - }, - { - "left": "issue:34928", - "right": "issue:36010", - "accept": false, - "reason": "FSDP activation-checkpointing mismatch and GenerationMixin import error are unrelated." - }, - { - "left": "issue:45237", - "right": "issue:46015", - "accept": false, - "reason": "GPT-OSS GPU failure and outdated RAG docs are unrelated." + "reason": "Phi autocast dtype mismatch and MPS histogram error are unrelated." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger kernel cross-entropy call issue and mask-shape IndexError are unrelated." + "reason": "Liger Kernel cross_entropy issue and mask/index shape mismatch are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The ZeRO-3/BERT duplicate pair is well supported by the nearly identical titles, and the torch-compatibility pair is presented as only plausible rather than certain. No soft-edge verdict appears overstated." + "feedback": "The summary is grounded in the packet: it correctly treats the cluster as mostly heterogeneous, while identifying the few clearly overlapping pairs (43596/43638, 45859/45902, 45800/46515). The negative soft-edge calls are generally conservative and supported by the titles, and the choice of 45237 as the broad representative is reasonable for the device/runtime theme." }, "evaluator_used": true, "retried": false, @@ -641,8 +611,8 @@ "issue:39290|issue:45823", "issue:36010|issue:39290", "issue:45464|issue:45685", - "issue:44841|issue:45084", "issue:43541|issue:45685", + "issue:44841|issue:45084", "issue:43541|issue:43828", "issue:43404|issue:43866", "issue:45507|issue:46869", @@ -651,160 +621,160 @@ "issue:43572|issue:45070", "issue:45800|issue:45823" ], - "prepared_review_unit_hash": "2262e91c2b2c139c1cd2d4c9b00df165b7f802dd664c5d8720f727f755f7e4e4", + "prepared_review_unit_hash": "47bcb7356cf97a5a065b79aa2e0846ce1cf5f656fea68f0f637670c817fd8053", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 42, - "serialized_chars": 40045, - "estimated_input_tokens": 10012, - "estimated_eval_tokens": 20280 + "serialized_chars": 40038, + "estimated_input_tokens": 10010, + "estimated_eval_tokens": 20276 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a grab bag of unrelated transformer/model-loading regressions. The only likely duplicate pair is the AutoTokenizer wrong-token-ID regression, where the broader multi-model report subsumes the Granite-specific one.", - "confidence": 0.88, - "canonical_issue_reason": "issue:45920 is the broadest and most representative report of the v5 AutoTokenizer wrong-token-ID regression across multiple model families; issue:45812 is a narrower Granite-only manifestation.", + "summary": "The set is mostly heterogeneous, but two soft edges look like true duplicate subsets: the DeepSpeed ZeRO-3 load-time IndexError reports and the v5 AutoTokenizer wrong-token-id regression reports.", + "confidence": 0.61, + "canonical_issue_reason": "issue:45920 is the best canonical issue for the tokenizer-regression subset because it describes the broad v5 AutoTokenizer wrong-token-ID bug across multiple model families, subsuming the narrower Granite-only report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 best captures the shared underlying bug and has the widest impact scope, making it the strongest canonical issue for this cluster.", + "best_issue_reason": "issue:45920 is the strongest overall issue anchor: it is broad, specific about the regression, and clearly actionable.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Different models and failures: corrupted checkpoint loading vs missing text encoder weights." + "reason": "Different models and different failure modes: corrupted checkpoint vs. missing weight loading in Sam3Model." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "Both are version-compatibility complaints, but they target different incompatibilities and code paths." + "reason": "Both are compatibility complaints, but they point to different breaks and different trigger conditions." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "PaddleOCR-VL/vLLM load failure is unrelated to USE_DISTRIBUTED=0 support." + "reason": "PaddleOCR-VL loading in vLLM is unrelated to the USE_DISTRIBUTED=0 PyTorch support regression." }, { "left": "issue:41762", "right": "issue:43596", - "accept": false, - "reason": "Both mention Zero-3 index errors, but one is Gemma3-specific loading and the other is a BertModel initialization failure." + "accept": true, + "reason": "Both report the same ZeRO-3 load/init path failing with an index-0 out-of-bounds error; model names differ, but the underlying bug appears shared." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3 config attribute error in vLLM is unrelated to Qwen3 auto device mapping cuda asserts." + "reason": "Gemma3 config attribute error and Qwen3 auto-device-map cuda assert are different code paths and symptoms." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Both are documentation issues, but they describe different docs and different behavior gaps." + "reason": "Both are docs issues, but they concern different API docs and do not share the same underlying bug." }, { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to the distributed-build regression." + "reason": "Float8 storage lookup failure is unrelated to the PyTorch distributed support regression." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "AMD GPU support for GPT-OSS-20B is unrelated to table QA crashes." + "reason": "AMD GPU support for GPT-OSS is unrelated to the table-question-answering crash." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Tokenizer detokenization bug and outdated RAG examples are unrelated." + "reason": "Incorrect detokenization and outdated RAG examples are unrelated issues." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross_entropy routing is unrelated to Qwen3-vl video StopIteration." + "reason": "Liger Kernel cross-entropy behavior and Qwen3-vl video StopIteration are different problems." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Device-mapping mismatch and chat-template video token mismatch are different failures." + "reason": "Gemma4 device-mismatch on auto device mapping is unrelated to Qwen3.5 video token/template mismatches." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Processor failure for Voxtral is unrelated to Gemma4 PLE device placement." + "reason": "Voxtral processor failure and Gemma4 device mismatch are distinct model-specific issues." }, { "left": "issue:43866", "right": "issue:46772", "accept": false, - "reason": "Checkpoint corruption and NVFP4 compatibility are different issues." + "reason": "Checkpoint corruption and NVFP4/ModelOpt compatibility are not the same bug." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint loading is unrelated to RAG docs." + "reason": "NemotronH checkpoint loading and RAG docs are unrelated." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Remote-code breakage and USE_DISTRIBUTED=0 support are different regressions." + "reason": "Remote-code breakage and PyTorch USE_DISTRIBUTED=0 support are different regressions." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Same underlying v5 AutoTokenizer wrong-token-ID regression; 45812 is a Granite-specific instance and 45920 is the broader multi-model report." + "reason": "These are the same v5 AutoTokenizer wrong-token-ID regression; 45920 is the broader report covering Granite and other affected model families." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Corrupted checkpoint vs model implementation cannot load checkpoints are not the same bug." + "reason": "Corrupted checkpoint vs. checkpoint loading failure for a different model family are not the same underlying bug." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error is unrelated to RAG documentation." + "reason": "Qwen NVFP4 runtime failure is unrelated to RAG documentation examples." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes Int8Params API mismatch is unrelated to the distributed build regression." + "reason": "bitsandbytes Int8Params incompatibility is unrelated to the distributed-build regression." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Different models and symptoms; both mention sliding-window-ish behavior but not the same concrete bug." + "reason": "Both mention sliding_window, but one is a missing config attribute in Gemma3/vLLM while the other is a Qwen3-MoE sliding_window issue; too different to merge." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are IndexErrors, but one is a mask-shape mismatch and the other is a Zero-3 loading failure." + "reason": "The mask-shape mismatch and the ZeRO-3 empty-index error are different IndexError classes with different contexts." }, { "left": "issue:43792", @@ -816,127 +786,127 @@ "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Unit-test load failure for GLM-4.7-Flash is not the same as Ovis2 checkpoint corruption." + "reason": "GLM-4.7-Flash loading in tests and Ovis2 checkpoint corruption are distinct model-specific issues." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch version compatibility and PretrainedConfig pydantic-field regression are different problems." + "reason": "Torch-version incompatibility and pydantic PretrainedConfig breakage are unrelated regressions." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "Tokenization/detokenization bug is unrelated to table QA crashes." + "reason": "Detokenization bug and table-question-answering crash are unrelated." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both involve Qwen3.5 chat-related failures, but one is a transformers chat crash and the other is streaming chat/completions failure." + "reason": "Both involve Qwen3.5 chat, but one is a transformers chat crash and the other is a streaming chat/completions API failure; not the same concrete bug." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "Two different Zero-3/version issues with different failure modes." + "reason": "ZeRO-3 index error and torch/transformers version incompatibility are different problems." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper load failure and PretrainedConfig pydantic regression are unrelated." + "reason": "Whisper loading and pydantic model breakage are unrelated." }, { "left": "issue:43541", "right": "issue:46869", "accept": false, - "reason": "Grouped_mm tracing failure and YOLO multi-GPU device-map failure are different code paths." + "reason": "Mixtral grouped_mm tracing failure and YOLO multi-GPU device-map failure are different issues." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "Qwen NVFP4 runtime error is unrelated to table QA." + "reason": "NVFP4 runtime failure and table QA crash are unrelated." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Gemma3 config attribute error and Gemma4 device mismatch are different bugs." + "reason": "Gemma3 sliding_window_pattern attribute error and Gemma4 device-mismatch are different model/path failures." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "Importing GenerationMixin and Gemma3TextConfig attribute access are unrelated." + "reason": "GenerationMixin import error and Gemma3TextConfig sliding_window_pattern error are unrelated." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Streaming chat API failure and MPS histogram op limitation are unrelated." + "reason": "Qwen3.5 streaming inference and MPS histogram_int support are unrelated." }, { - "left": "issue:44841", - "right": "issue:45084", + "left": "issue:43541", + "right": "issue:45685", "accept": false, - "reason": "Processor failure for Voxtral and template-node compilation error are different issues." + "reason": "Mixtral tracing failure and MPS histogram_int error are unrelated." }, { - "left": "issue:43541", - "right": "issue:45685", + "left": "issue:44841", + "right": "issue:45084", "accept": false, - "reason": "Torch-dynamo grouped_mm tracing failure is unrelated to the MPS histogram Int limitation." + "reason": "Voxtral processor failure and template-node compilation error are different bugs." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Both are model/runtime numeric issues, but they affect different models and execution paths." + "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different MoE-related symptoms, not the same code-path bug." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Untied lm_head weights in Mistral3 are unrelated to Ovis2 checkpoint corruption." + "reason": "Mistral3 lm_head tying issue and Ovis2 checkpoint corruption are unrelated." }, { "left": "issue:45507", "right": "issue:46869", "accept": false, - "reason": "GraniteMoEHybrid invalid method call and YOLO multi-GPU device-map failure are different bugs." + "reason": "GraniteMoEHybrid invalid method and YOLO multi-GPU auto device-map failure are different model-specific issues." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are IndexErrors, but the tensor-shape mismatch and Zero-3 loading failure are not the same underlying bug." + "reason": "Both are IndexErrors, but the mask/tensor shape mismatch is not the same as the ZeRO-3 empty-index load failure." }, { "left": "issue:45399", "right": "issue:46661", "accept": false, - "reason": "flash-attn2 fallback gating and remote-code config_class lookup are unrelated loading regressions." + "reason": "Flash-attn fallback blocking and remote-code config_class errors are unrelated." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig and pydantic PretrainedConfig breakage are different config regressions." + "reason": "Missing pad_token_idx in StableLmConfig and pydantic PretrainedConfig breakage are different regressions." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/version incompatibility and Gemma4 device mismatch are unrelated." + "reason": "Torch/transformers compatibility and Gemma4 device mismatch are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative: most items are indeed unrelated regression reports, and the 45812 \u2194 45920 pairing is well-supported by the titles and scope difference. I don\u2019t see an overstatement of evidence in the soft-edge verdicts." + "feedback": "The summary is grounded in the packet and stays conservative overall. The two accepted soft edges are well-supported by the titles: the ZeRO-3 index-out-of-bounds reports share the same failure shape, and 45920 clearly subsumes the Granite-only AutoTokenizer regression in 45812. No evidence looks overstated." }, "evaluator_used": true, "retried": false, @@ -1032,27 +1002,27 @@ "issue:43404|issue:45005", "issue:42915|issue:45399", "issue:43873|issue:45005", - "issue:44360|issue:44512", "issue:44530|issue:45005", + "issue:44360|issue:44512", "issue:44492|issue:44512", "issue:44863|issue:45005", "issue:46489|issue:46491", "issue:45183|issue:46463" ], - "prepared_review_unit_hash": "aebe31d42510554b07ce0bb8004a2fafbeac61c8bd1c3b1264b38ee1a0a8c106", + "prepared_review_unit_hash": "7053eab7b8448d871516f63d8695caa83f41a47f94de8eebad75a15bf3777738", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36892, + "serialized_chars": 36891, "estimated_input_tokens": 9223, "estimated_eval_tokens": 18702 }, @@ -1061,90 +1031,90 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous, with only one clear duplicate-style pair: the Qwen video-input regression report is a broader parent of the later Qwen3.5-9B apply_chat_template failure. The rest are mostly related only at a subsystem/theme level, not the same concrete bug.", + "summary": "Mostly a heterogeneous bucket of unrelated issues; only a few soft edges look like real duplicates. The strongest duplicate-like groups are the Qwen3.5 video-input regression, the remote-code/custom-module loading collision, and the input validation error-message bug. Most other pairs are different model families or different failure modes.", "confidence": 0.74, - "canonical_issue_reason": "issue:44479 is the best canonical issue because it describes the broader video-input regression across multiple Qwen VL/3.5 variants, and issue:46294 reads like a narrower instance of that same failure mode.", + "canonical_issue_reason": "issue:45183 is the broadest root-cause report in the set and cleanly subsumes the T5-specific validation symptom in issue:46463.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 is the strongest issue to keep as the representative bug report: it has the broadest scope, matches the concrete symptom pattern, and is backed by discussion that spans multiple affected model variants.", + "best_issue_reason": "issue:45183 is the best standalone representative because it describes the underlying validation bug, not just one model-family manifestation.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "Different failures: PaddleOCR-VL load/version compatibility vs GGUF gpt-oss architecture support." + "reason": "Different problem domains: PaddleOCR-VL load failure vs gpt-oss GGUF architecture support." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Both mention shape-related errors, but one is an index/mask mismatch and the other is a trainer/input-embedding unpacking failure; not the same bug path." + "reason": "Both mention shape-related failures, but the code paths and symptoms differ (mask indexing vs TRL embedding unpacking)." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "OOM under int4 quantization is unrelated to tied-weight problems in translation models." + "reason": "Unrelated: int4 quantization memory/OOM vs tied-weights behavior on translation models." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Both involve MoE models, but the concrete failures differ: torchdynamo grouped_mm tracing vs tensor-parallel support." + "reason": "Both involve MoE models, but one is a grouped_mm tracing error and the other is a tensor-parallel support issue." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch on Phi-tiny-MoE is a different code path from Qwen3.5 streaming chat/completions failures." + "reason": "Different failures: autocast dtype mismatch vs chat/completions streaming inference on Qwen3.5-0.8B." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "Unrelated model families and symptoms: GGUF architecture support vs streaming inference API failure." + "reason": "No concrete overlap beyond both being Qwen/gpt-oss-era model issues." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "One is an init_empty_weights argument-compatibility error; the other is a pydantic/PretrainedConfig field regression." + "reason": "Both are v5 regressions, but one is init_empty_weights argument handling and the other is a PretrainedConfig/pydantic field breakage." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Sliding-window behavior on Qwen3-MoE is unrelated to the PretrainedConfig/pydantic breakage." + "reason": "Different bug classes: sliding_window handling vs config serialization/type issues." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are IndexError reports, but they arise from different mechanisms: mask-length mismatch vs deepspeed zero3/BertModel empty-shape access." + "reason": "Mask-length mismatch and zero-sized index out-of-bounds are distinct runtime errors in different paths." }, { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Image-preprocessor dtype mismatch on Mistral-3 is not the same as Qwen video token/feature count mismatch." + "reason": "Both are multimodal input mismatches, but image preprocessor dtype mismatch and video token/feature mismatch are not the same bug." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote-code model breakage and torch/transformers incompatibility are too broad and not the same concrete defect." + "reason": "A broad remote_code regression and a torch/transformers version incompatibility are different root causes." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Both involve siglip2, but one reports degraded embedding quality and the other nonsensical pipeline outputs; not enough evidence of one underlying bug." + "reason": "Same model family, but one is a quality/performance complaint while the other is an inference correctness bug; not clearly the same code path." }, { "left": "issue:44479", "right": "issue:46294", "accept": true, - "reason": "Same underlying video-input regression: 46294 is a specific Qwen3.5-9B manifestation of the broader issue described in 44479." + "reason": "46294 looks like a specific Qwen3.5 video-input manifestation of the broader video regression described in 44479." }, { "left": "issue:46468", @@ -1156,109 +1126,109 @@ "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Composite-config unwrapping and uninitialized rotary embedding memory are different model-init bugs." + "reason": "Different model/config problems: composite config unwrapping vs uninitialized rotary embedding memory after materialization." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "Deepspeed zero3/BertModel index error is unrelated to resuming training with local trust_remote_code models." + "reason": "Different areas and symptoms: Deepspeed zero3/BertModel index error vs trust_remote_code resume behavior." }, { "left": "issue:46640", "right": "issue:46661", "accept": false, - "reason": "Flash Attention 2 integration issues are not the same as the remote-code model_class/config_class attribute error." + "reason": "Flash Attention 2 fallback logic and remote-code model_class config_class errors are unrelated." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Different multimodal validation failures: image preprocessing dtype mismatch vs SFT trainer embedding unpacking." + "reason": "Different multimodal/text paths; image preprocessor dtype mismatch is not the same as TRL embedding unpacking failure." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP 2-D weight error is unrelated to the Qwen3.5 input-embedding trainer failure." + "reason": "FSDP weight dimensionality error and Qwen3.5 embedding unpacking are unrelated." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "Perplexity regression on gpt-oss-20b and GLM5 are unrelated model-specific evaluation issues." + "reason": "High perplexity on gpt-oss-20b and GLM5 are separate model-specific reports." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Jupyter/custom-model init breakage is unrelated to fallback-to-flash-attn2 gating." + "reason": "Custom-model notebook initialization and flash-attn2 fallback gating are different bugs." }, { "left": "issue:45632", "right": "issue:45698", - "accept": false, - "reason": "Both are trust_remote_code/load-cache themed, but one is a leaf-directory cache collision and the other is wrong custom-module loading after save_pretrained; not clearly the same concrete bug." + "accept": true, + "reason": "Both point to remote-code/custom-module loading selecting the wrong local module; the latter is a concrete symptom of the cache/path collision problem." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both touch tied weights, but one is a specific missing lm_head tie in Mistral3 while the other is a broader v5 translation-model tied-weight regression." + "reason": "Tied lm_head weights in Mistral3 and translation-model tied-weights behavior are different model families and failures." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "FineGrainedFP8Config failure on Qwen3MoE is unrelated to flash-attn2 fallback gating." + "reason": "FineGrainedFP8Config failure and flash-attn2 fallback blocking are unrelated." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior and tied-weight translation-model issues are different problems." + "reason": "Quantization/offloading behavior and tied-weight translation issues are unrelated." }, { - "left": "issue:44360", - "right": "issue:44512", + "left": "issue:44530", + "right": "issue:45005", "accept": false, - "reason": "A DSA indexer/debug discussion is unrelated to a docs typo." + "reason": "PagedAttentionCache linear_attention crash is not the same as tied-weights handling." }, { - "left": "issue:44530", - "right": "issue:45005", + "left": "issue:44360", + "right": "issue:44512", "accept": false, - "reason": "PagedAttentionCache linear_attention crash on Qwen3.5 is unrelated to tied-weight translation model behavior." + "reason": "DSA indexer/ReLU discussion and docs for removed run command are unrelated." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "Cache-strategy typo and removed-run-command docs are different documentation issues." + "reason": "A documentation typo and a removed-command docs issue are not duplicates." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading failure is unrelated to tied weights in translation models." + "reason": "NemotronH checkpoint loading and tied weights in translation models are different bugs." }, { "left": "issue:46489", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different tokenizers/models and different round-trip behaviors; not one underlying bug." + "reason": "Different tokenizer regressions affecting different models and different round-trip behaviors." }, { "left": "issue:45183", "right": "issue:46463", - "accept": false, - "reason": "Both are validation/error-message bugs, but 45183 is the generic input_ids/inputs_embeds XOR issue while 46463 is specifically T5 decoder-input validation; too little evidence they are the same bug." + "accept": true, + "reason": "46463 is a T5-family-specific instance of the broader input_ids/inputs_embeds validation/error-message bug in 45183." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative on the soft pairs. The 44479 \u2194 46294 linkage is supported by the titles and looks like the only clearly mergeable duplicate-style pair; the other false verdicts are cautious and not overstated." + "feedback": "The summary is grounded in the packet and stays appropriately cautious. The three accepted soft edges are plausible duplicates/specific manifestations from the titles and descriptions, and the rejected pairs are conservatively framed as different bug classes or model families. I don\u2019t see an overstatement that would require rejection." }, "evaluator_used": true, "retried": false, @@ -1360,224 +1330,224 @@ "issue:43782|issue:44960", "issue:43824|issue:43872" ], - "prepared_review_unit_hash": "830ffaaaaab6006509c74d6a680576f58bca275a089251bfd86c38a5a80bac8f", + "prepared_review_unit_hash": "a7db8f174bdefa87da636906d175869ca00b0c36e2d6ff6eb9b58c0fa6139798", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36793, - "estimated_input_tokens": 9199, - "estimated_eval_tokens": 18654 + "serialized_chars": 36789, + "estimated_input_tokens": 9198, + "estimated_eval_tokens": 18652 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose set of Transformers v5 regression reports with several near-neighbors, but the soft pairs do not clear a conservative duplicate bar. The non-persistent-buffer reports are the strongest family signal, while the tokenizer/model-loading/quantization items mostly look related by subsystem rather than the same concrete bug.", - "confidence": 0.58, - "canonical_issue_reason": "issue:43950 is the best representative anchor for the cluster because it captures the clearest reusable regression family (non-persistent buffer corruption in v5), but it should be treated as a family representative rather than a confirmed duplicate of the surrounding reports.", + "summary": "This cluster is mostly heterogeneous, with a few true duplicate chains around specific regressions. The strongest duplicate set is the Transformers 5.x non-persistent-buffer regression, plus a separate exact match around Gemma4 quantized embedders and another around cache sequence-length return types.", + "confidence": 0.78, + "canonical_issue_reason": "issue:43950 is the best umbrella issue: it states the broad non-persistent-buffer regression clearly and covers the closely related manifestations in issue:44534 and issue:45902.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 is the best global issue representative because it states the broadest and most central v5 regression pattern in this set, even though many neighboring issues are only loosely related and should remain separate.", + "best_issue_reason": "issue:43950 is the most representative standalone issue in the cluster because it captures the shared underlying bug rather than one narrow symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Both are DeepSeek/Qwen-adjacent loading/tokenizer failures, but they involve different models and different failure modes." + "reason": "Both are model/tokenizer loading failures, but they involve different models and different failure modes; not the same bug." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different model families and different concrete bugs; TRL embedding unpacking is not the same as the GraniteMoE invalid-method issue." + "reason": "Unrelated problems: one is TRL SFT unpacking, the other is an invalid method call in a different model class." }, { "left": "issue:43950", "right": "issue:44534", - "accept": false, - "reason": "Both describe non-persistent-buffer corruption in v5, but the wording suggests a related manifestation rather than a confidently identical report." + "accept": true, + "reason": "Same underlying v5 non-persistent-buffer regression; 44534 is essentially the same symptom as 43950." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Docs-example/processor mismatch and quantization errors are not the same as the SigLIP2 embedding-quality regression." + "reason": "Both mention SigLIP2, but one is a documentation/example mismatch and the other is a quality discrepancy; not the same concrete bug." }, { "left": "issue:43644", "right": "issue:43950", "accept": false, - "reason": "These are related v5 buffer-loading regressions, but the evidence is not strong enough to treat them as one confirmed duplicate." + "reason": "Different areas of the codebase: remote-code config loading vs buffer materialization regression." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Custom trainer signature mismatch and DeepSpeed ZeRO-3 loading IndexError are different code paths and different failures." + "reason": "Different training/loading failures with no shared code path indicated." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Both touch quantization-related areas, but they are distinct incompatibilities with different error classes and causes." + "reason": "Different quantization/storage incompatibilities; the titles do not indicate the same underlying fix." }, { "left": "issue:46899", "right": "issue:46932", - "accept": false, - "reason": "The titles point to the same Gemma4 quantized embedder area, but the packet does not provide enough certainty to merge them as a confirmed duplicate." + "accept": true, + "reason": "46932 explicitly describes the same quantized vision/audio dtype-casting bug as 46899." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Autocast dtype mismatch and Qwen3.5 3D position_ids / SDPA behavior are different bugs, even if both appear in generation paths." + "reason": "Both are numerical/runtime issues, but they are different models and different mechanisms; not a duplicate." }, { "left": "issue:45920", "right": "issue:46489", "accept": false, - "reason": "The tokenizer regressions are related, but one is a broad multi-model v5 issue and the other is a DeepSeek-Coder-specific symptom." + "reason": "Both are tokenizer regressions in v5, but the titles point to different affected models and a broader vs specific failure; not enough to treat as the same bug." }, { "left": "issue:45987", "right": "issue:46858", - "accept": false, - "reason": "They are close on the API surface, but the backend behavior and downstream breakage are described differently enough to keep separate." + "accept": true, + "reason": "Same cache-length contract bug: get_seq_length is returning a tensor instead of an int/device scalar, breaking downstream arithmetic." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "A model-quality regression and a generic tokenization-change report are not the same concrete bug." + "reason": "One is a SigLIP2 embedding-quality complaint, the other is a broad tokenization-change report; unrelated." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both involve SigLIP2, but zero probabilities and degraded text embeddings are distinct symptoms and not clearly one defect." + "reason": "Both concern SigLIP2 outputs, but one is zero probabilities and the other is lower text embedding quality; not clearly the same root cause." }, { "left": "issue:43950", "right": "issue:45902", - "accept": false, - "reason": "Both fit the non-persistent-buffer family, but the meta->to_empty(cuda) report is a specific manifestation rather than a certain duplicate." + "accept": true, + "reason": "Both describe non-persistent buffers becoming junk/uninitialized after loading or materialization in Transformers 5.x." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Bad AutoTokenizer error messaging and Molmo tied-weights failure are different models and different root causes." + "reason": "Different model-loading errors with no shared failure mode." }, { "left": "issue:43054", "right": "issue:43493", "accept": false, - "reason": "These are both SigLIP2-related, but the reports focus on different aspects of behavior and implementation mismatch." + "reason": "Both are SigLIP2-related, but one is a generic performance comparison while the other is a specific HF-vs-JAX implementation discrepancy." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Both are Qwen3.5 issues, but flash-attn position handling and CPU-offload rotary embedding incompatibility are different bugs." + "reason": "Different Qwen3.5 issues: FlashAttention misinterpreting 3D position_ids vs CPU-offload incompatibility in rotary embeddings." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Gradient checkpointing/LoRA behavior and int4 OOM are both memory-adjacent, but not the same code-path problem." + "reason": "Both mention memory, but one is gradient checkpointing/LoRA behavior and the other is int4 quantization OOM; different bugs." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Gemma4 FSDP2 KeyError and non-persistent buffer corruption are unrelated failures." + "reason": "Gemma4 FSDP2 shared_kv_states KeyError is unrelated to the non-persistent-buffer regression." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise is not the same as an inability to load a specific model." + "reason": "One is a noisy-loading UX issue, the other is a model load failure; not duplicates." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Tokenizer detokenization and flash_attn import failure are unrelated bugs." + "reason": "DeepSeek-Coder detokenization and a flash_attn import error are unrelated." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "They are both load failures, but for different models and different failure causes." + "reason": "Different model-loading failures and different components." }, { "left": "issue:24643", "right": "issue:28530", "accept": false, - "reason": "DeepSpeed weight-shape training error and early-stopping metric selection are unrelated." + "reason": "Training with DeepSpeed and early-stopping metric handling are unrelated issues." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Both mention memory, but one is checkpointing/LoRA behavior and the other is Qwen3 MoE training memory usage." + "reason": "Different memory-related regressions; no shared code path implied by the titles." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Remote-code loading breakage is too broad to equate with missing fields in AutoConfig objects." + "reason": "Both mention remote code, but one is missing fields in AutoConfig and the other is a broad breakage report; too little to merge as the same bug." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Logging noise and a model-loading failure do not share the same underlying bug." + "reason": "Different symptoms and likely different model-loading paths." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Gemma3 ZeRO-3 loading and Qwen3Moe FineGrainedFP8 failure are distinct model/runtime issues." + "reason": "Completely different models and failures: ZeRO-3 loading vs FineGrainedFP8Config." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading regression and flash-attn fallback logic are different subsystems and different failures." + "reason": "Tokenizer loading and flash-attn fallback are unrelated." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight-only loading and GLM5 are not the same bug." + "reason": "Different model families and different error classes." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Missing model class import and bitsandbytes Int8Params incompatibility are unrelated failures." + "reason": "Import/export symbol availability and bitsandbytes parameter incompatibility are unrelated." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster summary is mostly plausible, but it overstates duplicate confidence. In the packet, every soft pair is marked deterministic_accept=false, yet the analyst flips several of them to accept=true (notably 43950\u219444534, 43644\u219443950, 43950\u219445902, 45920\u219446489, and 46899\u219446932). That is not conservative enough for this evaluator. The safest read is that these are related regression families or near-neighbors, not confirmed duplicates, so the canonical/best-issue language should be softened accordingly." + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative on the soft-edge pairs. The cluster summary correctly frames the set as mostly heterogeneous with a few duplicate chains, and the selected canonical issue/pairing around issue:43950 is well supported by the titles. I did not see any overstatement that would require rejection." }, "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": false, + "retried": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -1672,208 +1642,226 @@ "issue:44479|issue:44560", "issue:41553|issue:44488" ], - "prepared_review_unit_hash": "7c44ff36bb25511792d57dd568ac093e69240a90f0c2a3d3fa7b882f9d72d0cd", + "prepared_review_unit_hash": "14ba2c3c840d4f85ff4d31a021e1c9297de5624b7eb005f72d68e8f99f2af5ef", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36189, - "estimated_input_tokens": 9048, - "estimated_eval_tokens": 18352 + "serialized_chars": 36193, + "estimated_input_tokens": 9049, + "estimated_eval_tokens": 18354 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "A few clear duplicate pairs appear, mostly around tokenizer/docs regressions and one shared training-loss bug; most other soft edges are only broad subsystem similarity and should stay separate.", - "confidence": 0.83, - "canonical_issue_reason": "issue:44291 is the clearest root-cause report: it names the unexpected `_is_hf_initialized` loading failure, gives the affected version range, and captures the generic bug better than the more specific manifestation in issue:43872.", + "summary": "This set is mostly heterogeneous, with only a few true duplicate pockets: v5 docs still referencing removed pipeline tasks, repeated mask/tensor shape-mismatch reports, and a likely repeated training-loss double-shift report. The rest share broad subsystems or model families but do not look like the same underlying bug.", + "confidence": 0.76, + "canonical_issue_reason": "No single canonical issue cleanly covers the whole set; the items split into multiple unrelated bug pockets, so a global canonical would be misleading.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44291 is the strongest canonical issue candidate because it is explicit, reproducible, and general enough to represent the `_is_hf_initialized` loading regression cleanly.", + "best_issue_reason": "issue:44509 is the clearest representative of the docs duplicate pocket because it explicitly names the removed v5 pipeline tasks and has the broadest, most actionable description.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "Different failures: MPS histogram/int support vs Gemma4 device-mapping forward mismatch." + "reason": "Different failures: MPS int histogram support vs device-mismatch on PLE forward." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different models and different load errors; only share a broad \"loading failed\" theme." + "reason": "Different models and different loading bugs; only the broad symptom is similar." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Tokenizer regex crash and flash-attn2 fallback gating are unrelated code paths." + "reason": "Unrelated tokenizer regex crash vs flash-attn fallback gating." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote-code loading regression vs a device mismatch in Gemma4; not the same bug." + "reason": "Both are model-loading regressions, but the concrete failure modes are different." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Different loader paths and symptoms: tokenizer mapping bug vs remote-code compatibility regression." + "reason": "Different code paths: tokenizer mapping edge case vs remote-code loading regression." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "Both involve Qwen/LoRA, but one is an incompatibility to use a saved model and the other is a warning about tied embeddings." + "reason": "Same model family, but one is a LoRA compatibility break and the other is a warning about tie_word_embeddings." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Autocast dtype mismatch is a different failure from the device-map mismatch." + "reason": "Different runtime error classes: autocast dtype mismatch vs device placement mismatch." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Remote-code breakage and Kimi tokenizer codec handling are separate problems." + "reason": "Both are tokenizer/model regressions, but not the same bug or fix." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer performance regressions, but in different APIs with different bottlenecks." + "reason": "Both are tokenizer performance issues, but they hit different conversion directions and code paths." }, { "left": "issue:45488", "right": "issue:46489", "accept": false, - "reason": "Same family, but one is a tokenizer pre-tokenizer override regression and the other is incorrect tokenization output; not enough evidence they're the same bug." + "reason": "Related tokenizer regressions, but different model families and not enough evidence of one shared bug." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Same docs cleanup: both report stale summarization/translation pipeline references after v5 removed those tasks." + "reason": "Same docs regression: v5 removed pipeline tasks, and both issues report stale docs referencing them." }, { "left": "issue:46894", "right": "issue:46897", "accept": true, - "reason": "Same underlying training-loss double-shift bug; the titles describe the same pattern on different models." + "reason": "Same training-loss double-shift pattern reported for different models; likely the same underlying loss bug." }, { "left": "issue:45759", "right": "issue:46661", "accept": false, - "reason": "Different AutoModel loading regressions: missing `text_config` unwrap vs missing `config_class` on remote-code models." + "reason": "Different remote-code/composite-config failures with different missing attributes." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Weight tying and non-persistent buffer corruption are distinct model-loading issues." + "reason": "Different problems: untied lm_head weights vs non-persistent buffer corruption." }, { "left": "issue:43872", "right": "issue:46869", "accept": false, - "reason": "Bitsandbytes `_is_hf_initialized` incompatibility is unrelated to YOLO multi-GPU device-map behavior." + "reason": "Different device/loading contexts; bitsandbytes init arg mismatch is not the YOLO multi-GPU issue." }, { "left": "issue:43872", "right": "issue:44291", - "accept": true, - "reason": "Same root cause: loading code passes `_is_hf_initialized` to constructors that do not accept it." + "accept": false, + "reason": "Same unexpected-argument theme, but different libraries and loading paths." }, { "left": "issue:45020", "right": "issue:46661", "accept": false, - "reason": "Both touch remote-code models, but one is a general breakage and the other is a specific `config_class` attribute error." + "reason": "Both involve remote-code models, but the reported AttributeError differs from the broader v5 loading breakage." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "`whisper-large-v2` runtime failure and `_torch_extract_fbank_features()` are too vague to treat as the same bug." + "reason": "Both involve audio models, but one is a general Whisper run failure and the other is a specific fbank-feature issue." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Both are DeepSpeed-related, but one is missing weight conversion and the other is gradient-averaging behavior." + "reason": "Different DeepSpeed problems: weight conversion/initialization vs gradient accumulation averaging." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Gradient-checkpointing ineffectiveness in LoRA is not the same as a missing `supports_gradient_checkpointing` flag on a remote-code model." + "reason": "Different checkpointing issues: ineffective gradient checkpointing vs missing capability flag." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Different errors and code paths: trainer `compute_loss` signature vs tensor mask shape mismatch." + "reason": "Completely different error classes: trainer API keyword mismatch vs tensor mask shape mismatch." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Different tokenizer/model loading failures; one is a bad error message, the other is an outright tokenizer load break." + "reason": "Both are tokenizer/model loading problems, but for different models and different root causes." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different models and different loading failures." + "reason": "Different model-loading failures for different architectures; no shared bug evidence." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both are Qwen ZeRO-3 load failures, but the model variants and reported symptoms differ too much to merge confidently." + "reason": "Same subsystem, but different Qwen variants and different missing-weight patterns." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "RLHF NaN/SDPA fallback problem is unrelated to the flash_attention `s_aux=None` crash." + "reason": "Different failures: NaN/SDPA math fallback vs flash-attn sink auxiliary crash." }, { "left": "issue:41093", "right": "issue:44805", "accept": true, - "reason": "Same exact mask/tensor shape-mismatch error signature, so these are duplicates." + "reason": "Same mask-shape mismatch error signature; these look like duplicate reports of the same indexing bug." }, { "left": "issue:44479", "right": "issue:44560", "accept": false, - "reason": "Both involve video input regressions, but the affected models and failure modes are different." + "reason": "Both are video-related regressions, but one is a broad v5 video-input break and the other is a StopIteration embedding error." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both are model loading problems, but for different models and different symptoms." + "reason": "Different tokenizer-loading issues for different models and failure modes." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model families and different checkpoint/loading failures." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 load failures, but the concrete affected models and missing-weight patterns differ." + }, + { + "left": "issue:45305", + "right": "issue:46894", + "accept": false, + "reason": "Different training bugs: GAS averaging under DeepSpeed vs a double-shift loss issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted pairs are well-supported by matching titles/root-cause descriptions, especially 43827\u219444509, 43872\u219444291, 41093\u219444805, and 46894\u219446897. The rejected pairs are also framed cautiously and do not overclaim equivalence from broad subsystem similarity alone." + "feedback": "The summary is grounded in the packet and stays conservative about the few duplicate pockets it identifies. The accepted soft-edge pairs are well supported by the titles/descriptions, and the rest are correctly treated as related-but-distinct rather than the same bug. No overstatement stands out." }, "evaluator_used": true, "retried": false, @@ -1979,20 +1967,20 @@ "issue:43638|issue:44661", "issue:43596|issue:44661" ], - "prepared_review_unit_hash": "91138c5952011e14f89ef01d1348a74bc75325176fe2d992a51a82a4c9a72f0d", + "prepared_review_unit_hash": "88f32c22a73e1c1ad918c01b423218d43f98b32f4ae0a25eadfd890d2c651799", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38070, + "serialized_chars": 38069, "estimated_input_tokens": 9518, "estimated_eval_tokens": 19292 }, @@ -2001,222 +1989,222 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a collection of unrelated Transformers v5 regression reports. I only see a few true duplicate pairs: custom model initialization, SigLIP2 bad outputs, Gemma4 4-bit multimodal quantization, and config-validation loading failures.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 43632 is the broadest root-cause-style report in the set: it frames a general Transformers v5 initialization regression that can explain several downstream loading/custom-model failures.", + "summary": "Mostly a mixed cluster of unrelated issues, with a few small duplicate-like subgroups: custom model initialization regressions in Transformers 5.x, Siglip2 wrong outputs, and Gemma4 4-bit multimodal dtype-cast failures.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43646 is the clearest broad statement of the custom model initialization regression, while issue:43645 is just a notebook-specific manifestation.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43957 is the strongest standalone user-facing report: it describes a concrete model-loading regression on the meta device, names affected models, and gives a clear symptom pattern.", + "best_issue_reason": "issue:43646 is the best representative issue in this set because it is general, concise, and directly captures the shared initialization bug surface.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Different bugs: DeepSpeed weight-shape training error vs. segmentation-map preprocessing failure." + "reason": "Different problems: DeepSpeed training weight-shape error vs image processor handling void segmentation maps." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Different models and failures: Qwen3.5 cache/group-type crash vs. NemotronH checkpoint loading." + "reason": "Both are model/runtime failures, but the model families and failure mechanisms differ (Qwen3.5 cache path vs NemotronH checkpoint loading)." }, { "left": "issue:44863", "right": "issue:46402", "accept": false, - "reason": "Both are loading-related, but one is checkpoint incompatibility and the other is missing text-encoder weights; not the same bug." + "reason": "Different loading bugs: NemotronH checkpoint compatibility vs Sam3 text encoder weights not being loaded." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Unrelated: GGUF gpt-oss support request vs. Phi-tiny-MoE autocast dtype mismatch." + "reason": "Unrelated model-specific issues: GGUF gpt-oss support vs Phi-tiny-MoE autocast dtype mismatch." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Both involve rotary embeddings, but they affect different models and distinct code paths." + "reason": "Both involve Qwen variants and RoPE/state handling, but they target different code paths and symptoms." }, { "left": "issue:43596", "right": "issue:46515", "accept": false, - "reason": "Unrelated: Deepspeed ZeRO3/Bert index error vs. torch 2.4.1 compatibility issue." + "reason": "Different classes of failure: DeepSpeed ZeRO3/Bert initialization vs a torch version incompatibility report." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Different topics: GGUF architecture support vs. Qwen3-MoE sliding-window behavior." + "reason": "Same broad model family, but one is GGUF architecture support and the other is sliding-window behavior; not the same bug." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "Both are audio-related, but one is batched track leakage and the other is empty-transcription handling." + "reason": "Different audio-generation bugs: track-to-track embedding leak vs empty-transcription handling after special-token alignment." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Same model family, but different problems: FSDP2 shared_kv_states rebuild vs. device-map PLE mismatch." + "reason": "Both concern Gemma4, but one is an FSDP2 shared_kv_states bug and the other is a device_map forward mismatch." }, { "left": "issue:43638", "right": "issue:46515", "accept": false, - "reason": "Different issues: ZeRO3/non-pretrained Bert index error vs. torch version incompatibility." + "reason": "Specific DeepSpeed ZeRO3/Bert index error is not the same as a generic torch version incompatibility report." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Loading/buffer regression vs. OCR long-input index error; not the same bug." + "reason": "Non-persistent buffer corruption is unrelated to Kosmos2.5 long-OCR index errors." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "Unrelated: GGUF support vs. missing Float8 storage type." + "reason": "GGUF gpt-oss support and float8 storage lookup errors are different loading/runtime issues." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Both are import/loading failures, but one is a missing class export and the other is meta-device materialization breakage." + "reason": "Importing a missing class and meta-device model loading failures are different code paths." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Different Qwen3-MoE FP8 failure vs. general meta-device loading regression." + "reason": "Qwen3Moe FP8 failure and meta-device loading failure are different bugs, even if both are Transformers v5 regressions." }, { "left": "issue:44910", "right": "issue:46858", "accept": false, - "reason": "Different cache bugs: packed-sequence position_ids misread vs. device-tensor seq-length arithmetic." + "reason": "Both touch cache/position handling, but one is Qwen3.5 flash-attention on 3D position_ids and the other is static-cache seq-length arithmetic on non-CUDA backends." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Both are multimodal, but one is dtype mismatch in image preprocessing and the other is video StopIteration." + "reason": "Different models and symptoms: Mistral-3 image preprocessor dtype mismatch vs Qwen3-vl-embedding video StopIteration." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same underlying regression: Transformers 5.0 custom model initialization breaks, with one report highlighting Jupyter and the other the general case." + "reason": "Same underlying custom-model initialization regression; the notebook-specific report is a narrower manifestation of the broader issue." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Different attention/compile failures on different models and code paths." + "reason": "Different attention/compile failures on different models and kernels; not the same concrete bug." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Unrelated: Qwen3-VL video input failure vs. MPS histogram not implemented for Int." + "reason": "Unrelated failures: Qwen3-vl video input iteration vs MPS histogram on Int in MoE code." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "Different areas: 3D parallel script failure vs. GGUF architecture support." + "reason": "Different subsystems and failures; sharing a theme of model support is too broad to merge." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "Unrelated import error vs. Qwen3-MoE FP8 loading failure." + "reason": "Importing GenerationMixin and Qwen3Moe FP8 failures are unrelated bugs." }, { "left": "issue:44683", "right": "issue:46515", "accept": false, - "reason": "Both mention torch versions, but one is compiled flex_attention on torch>=2.9 and the other is torch 2.4.1 incompatibility." + "reason": "A torch>=2.9 flex_attention compilation regression is not the same as a torch 2.4.1 incompatibility report." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Same model family, but different bugs: video StopIteration vs. TRL SFT embedding unpacking." + "reason": "Both involve Qwen/VL embeddings, but one is video input StopIteration and the other is TRL SFT unpacking failure." }, { "left": "issue:45812", "right": "issue:46489", "accept": false, - "reason": "Both are tokenizer regressions, but they target different tokenizer implementations and model families." + "reason": "Both are tokenizer regressions, but they affect different model families and likely different mapping tables/code paths." }, { "left": "issue:45823", "right": "issue:46515", "accept": false, - "reason": "Different failures: Gemma4 device-map mismatch vs. generic torch 2.4.1 incompatibility." + "reason": "Gemma4 device-mapping bug and torch 2.4.1 incompatibility are unrelated." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Unrelated config-mapping bug vs. Gemma4 forward device mismatch." + "reason": "Tokenizer-mapping validation failure is unrelated to Gemma4 PLE device mismatch." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Same underlying SigLIP2 model-output regression: zero/garbage probabilities and nonsensical AutoModel/pipeline results." + "reason": "Same model and same wrong-output symptom for siglip2-base-patch16-224; these look like duplicate reports of the same bug." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Different shape/indexing errors on different models and inputs." + "reason": "StopIteration in Qwen3-vl video handling is unrelated to a tensor-mask shape mismatch error." }, { "left": "issue:45672", "right": "issue:46899", "accept": true, - "reason": "Same Gemma4 4-bit multimodal quantization bug: unsafe dtype casting in audio/vision embedders." + "reason": "Both point to the same Gemma4 4-bit multimodal dtype-casting problem affecting audio/vision embedders." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "General v5 init-flag regression vs. OCR long-input index error." + "reason": "Custom-model init flag regression and Kosmos2.5 long-OCR index errors are different bugs." }, { "left": "issue:44877", "right": "issue:45030", - "accept": true, - "reason": "Same class of regression: stricter config validation blocks model config loading in both reports." + "accept": false, + "reason": "Both are config-validation regressions, but they are on different model configs and do not look like one shared bug." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Different loading/compatibility bugs: Qwen3-MoE LoRA reuse across versions vs. BitNet packed-weight unpacking." + "reason": "LoRA/Qwen3_moe loading on hf4.x and BitNet packed-weight loading are unrelated." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Unrelated ZeRO3 Bert failure vs. tokenizer-mapping-name add-new-model bug." + "reason": "DeepSpeed/Bert initialization index error and tokenizer-mapping-name validation are unrelated." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Different code paths and symptoms: ZeRO3 index error vs. tokenizer mapping generation issue." + "reason": "ZeRO3/Bert init failure and tokenizer mapping validation are unrelated bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall. The accepted pairs are plausible duplicates from the packet, and the rejected pairs are mostly conservative distinctions rather than overclaims. The summary also stays within the evidence and does not overstate certainty." + "feedback": "The summary is grounded in the packet and stays conservative: it identifies the broad custom-model initialization cluster while treating the other pairs as only similar or unrelated. The accepted duplicate-like pairs are well supported by the titles and the reasons avoid overstating evidence." }, "evaluator_used": true, "retried": false, @@ -2302,8 +2290,8 @@ "issue:44488|issue:44991", "issue:45507|issue:46468", "issue:43644|issue:43645", - "issue:39692|issue:43994", "issue:43257|issue:43854", + "issue:39692|issue:43994", "issue:43856|issue:43873", "issue:42491|issue:45292", "issue:43856|issue:44387", @@ -2318,20 +2306,20 @@ "issue:45276|issue:45335", "issue:44910|issue:45588" ], - "prepared_review_unit_hash": "bb55fcefc99ba3fed47ebdcf77dedffb4acea438870924bc99bc7a2d740b73cd", + "prepared_review_unit_hash": "19124452ae96b7e598086f86b40a037ed7a6784700f152e914f6586b1fe1440f", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36656, + "serialized_chars": 36655, "estimated_input_tokens": 9164, "estimated_eval_tokens": 18584 }, @@ -2340,201 +2328,201 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster contains a few real duplicate subclusters, but most pairs are distinct bugs. The strongest true duplicates are the v5 non-persistent-buffer regression (43644 \u2194 44534), the ScandiBERT/sleng-bert loading failure (44451 \u2194 44488), and the training-loss double-shift bug (46897 \u2194 46901). The SigLIP2 pair (38175 \u2194 39692) should stay rejected: it is only a same-model-family resemblance, not the same underlying bug.", - "confidence": 0.82, - "canonical_issue_reason": "43644 is the cleanest representative of a confirmed duplicate subcluster: it states a concrete v5 regression and has a near-identical companion in 44534 with the same symptom and fix path.", + "summary": "This cluster is mostly a mix of unrelated model-specific regressions and loading/quantization issues. The only clear duplicate pair is the Transformers v5 non-persistent-buffer junk bug; the rest share broad themes like model loading, tokenizer behavior, or quantization but do not look like the same underlying defect.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43644 is the best canonical issue because it states the core regression most clearly and broadly; 44534 is essentially the same bug with a slightly different title.", "canonical_pr_reason": null, - "best_issue_reason": "43644 is the best global representative because it cleanly names the underlying bug in a broad, clearly duplicated regression cluster.", + "best_issue_reason": "43644 is the strongest representative of the duplicate subcluster and the most general statement of the shared bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Different bugs: quantization/offload loading failure vs Strix Halo mmap/OOM behavior." + "reason": "Different problems: gpu+cpu offload/dequantized loading vs Strix Halo mmap/OOM." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Different failure modes: tokenizer-mapping edge case vs tensor/mask shape mismatch." + "reason": "Unrelated failure modes: tokenizer mapping-name registration vs mask/tensor shape mismatch." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Same model family, but different concrete problems: inference output regression vs documentation/example plus quantization errors." + "reason": "Different SigLIP2 issues: wrong probabilities vs documentation example/model-processor errors." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Same v5 regression: non-persistent buffers being filled with junk." + "reason": "Same underlying Transformers v5 regression about non-persistent buffers being filled with junk; title wording differs only slightly." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Both are Qwen3/Qwen3-MoE related, but the concrete failures differ: device-map assert vs FP8 config failure." + "reason": "Both are Qwen3-family, but one is auto device mapping CUDA assert and the other is FineGrainedFP8Config failure; not the same bug." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Related resize/load themes, but different models and broken paths." + "reason": "Different code paths and symptoms: LoRA checkpoint compatibility across branches vs resize_token_embeddings not updating Gemma4 embeddings." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Unrelated problems: Qwen3-MoE FP8 training vs Strix Halo mmap/OOM." + "reason": "Different model/setting: Qwen3Moe FP8 failure vs mmap-related OOM on Strix Halo." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Different model families and different flash-attention failure reports." + "reason": "Different models and triggers: GLM5 issue vs Qwen3.5 flash-attention generation bug." }, { "left": "issue:44451", "right": "issue:44488", - "accept": true, - "reason": "Same model-loading regression affecting two checkpoints; both report the model no longer loads correctly." + "accept": false, + "reason": "Both are model-loading regressions, but they concern different repositories and no shared concrete code-path is evident." }, { "left": "issue:44863", "right": "issue:46846", "accept": false, - "reason": "NemotronH checkpoint loading is a different issue from Qwen3.5 TP-plan OOM/shape errors." + "reason": "Different bugs: NemotronH checkpoint loading vs Qwen3.5 tensor-parallel plan missing linear_attn entries." }, { "left": "issue:46459", "right": "issue:46512", "accept": false, - "reason": "Both are dtype-related, but one is composite-checkpoint loading and the other is from_config dtype propagation." + "reason": "Both involve dtype handling, but one is loading composite checkpoints and the other is from_config stale config metadata." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 regressions: ZeRO-3 weight loading vs incorrect visual-encoder save_pretrained keys." + "reason": "Different issues: ZeRO-3 weight loading for language_model vs incorrect visual encoder keys on save_pretrained." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are loading problems, but for different models and likely different tokenizer/model-conversion causes." + "reason": "Both are tokenizer/model loading regressions, but they affect different model families and likely different root causes." }, { "left": "issue:45507", "right": "issue:46468", "accept": false, - "reason": "Different model families and no shared concrete failure path." + "reason": "Different models and failures: invalid method call in GraniteMoEHybrid vs broken Nemotron example." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "One is a buffer-initialization regression; the other is a Jupyter/custom-model initialization problem." + "reason": "One is a buffer-initialization regression; the other is a Jupyter custom-model initialization breakage." }, { - "left": "issue:39692", - "right": "issue:43994", + "left": "issue:43257", + "right": "issue:43854", "accept": false, - "reason": "Same SigLIP2 family, but the first is docs/example errors while the second is an inference regression." + "reason": "Different symptoms: Qwen3 MoE weights conversion with accelerate+deepspeed vs GLM-4.7-Flash unit-test loading." }, { - "left": "issue:43257", - "right": "issue:43854", + "left": "issue:39692", + "right": "issue:43994", "accept": false, - "reason": "Different issues: accelerate+deepspeed MoE weight conversion vs a GLM-4.7 model-loading test failure." + "reason": "Similar model family, but one is docs example errors and the other is nonsensical outputs with AutoModel/pipeline." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both mention memory/offload, but the concrete problems differ: MoE training memory usage vs quantization offloading behavior." + "reason": "Both mention memory/quantization, but one is Qwen3 MoE training memory use and the other is offloading behavior with quantization." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "Related resize/save concerns, but not the same model or the same broken embedding path." + "reason": "Different concerns: checkpoint compatibility across branches vs resize_token_embeddings not propagating to output embeddings." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Different performance regressions: MoE training memory usage vs int4 quantization CUDA reserved memory/OOM." + "reason": "Both are memory-related, but one is training memory usage and the other is increased reserved CUDA memory under int4 quantization." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different model-loading failures for different models and different root causes." + "reason": "Different loading failures for different models; no clear shared code-path." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer checkpoint-resume/max_steps logic is unrelated to SGD argument handling." + "reason": "Trainer checkpoint-resume step math vs optimizer SGD args being ignored are unrelated." }, { "left": "issue:42491", "right": "issue:45335", "accept": false, - "reason": "Both involve embedding resizing, but the affected models and broken paths differ." + "reason": "Both mention embedding/token resizing, but they target different model internals and do not read like the same bug." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "Different bugs: GLM5 issue vs sinq save/load corruption." + "reason": "Different issues: GLM5 model behavior vs sinq quantized save/load breakage." }, { "left": "issue:46897", "right": "issue:46901", - "accept": true, - "reason": "Same training-loss double-shift bug: both train against shifted labels instead of the intended labels tensor." + "accept": false, + "reason": "Same high-level pattern (double-shift training loss), but different model implementations and no evidence one fix would cover both." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different tokenizers/models and are not the same concrete fix." + "reason": "Both are tokenizer regressions, but they affect different model families and different failure mechanisms." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Different crashes in different components: chat generation vs serve/processor attribute error." + "reason": "Different runtime failures: Qwen3.5 chat crash vs Gemma4Processor missing _tokenizer in transformers serve." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Both are save/load related, but extra_state handling and init_empty_weights argument handling are separate bugs." + "reason": "Both involve save/load behavior, but extra_state handling is distinct from the _is_hf_initialized init_empty_weights TypeError." }, { "left": "issue:45276", "right": "issue:45335", "accept": false, - "reason": "Similar resize_token_embeddings symptom, but different model families and different missing targets." + "reason": "Related resize_token_embeddings symptom, but different model families and embedding destinations; not enough evidence of one shared bug." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Different flash-attention crashes with different triggering conditions and code paths." + "reason": "Both are flash-attention related, but one is a packed-sequence misinterpretation and the other is an s_aux=None AttributeError." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The output is mostly grounded, but it overstates one subcluster: the summary calls the SigLIP2 pair a \u201cclearest true duplicate,\u201d while the soft verdict explicitly rejects 38175 \u2194 39692 and the packet only supports a loose same-family resemblance, not the same bug. Keep the summary aligned with the conservative pair decisions and avoid promoting rejected pairs to true duplicate clusters." + "accept": true, + "feedback": "The summary is grounded in the packet, and the only accepted duplicate pair (43644/44534) is strongly supported by near-identical titles. The remaining soft-edge rejections are conservative and do not overclaim shared root causes." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -2629,23 +2617,23 @@ "issue:45588|issue:45663", "issue:43854|issue:45216", "issue:45292|issue:45335", - "issue:46894|issue:46901", - "issue:44402|issue:44451" + "issue:44402|issue:44451", + "issue:46894|issue:46901" ], - "prepared_review_unit_hash": "83759f47204baa9d6b689c25e6e1cdc083a71caad814eadc02abc0e771531ef5", + "prepared_review_unit_hash": "dc4143c817a652e1916b30526e7c36826335c6cdaf2ff8654740b78bb0f37d52", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36140, + "serialized_chars": 36138, "estimated_input_tokens": 9035, "estimated_eval_tokens": 18326 }, @@ -2654,201 +2642,201 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous overall, but it does contain multiple coherent duplicate subfamilies: the main `resize_token_embeddings` propagation bug, a smaller training-loss double-shift pair, and a near-match around `lm_head` tying. The earlier writeup was too absolute in calling the resize family the only meaningful duplicate pattern.", - "confidence": 0.82, - "canonical_issue_reason": "`issue:45276` is still the best canonical issue for the largest and clearest family: `resize_token_embeddings` not propagating updated embeddings to downstream/output embeddings. It is the broadest, most concrete representative of that subcluster, with `issue:45292` and `issue:45335` as direct variants.", + "summary": "This cluster is mostly heterogeneous issue reports with one clear duplicate pair around `resize_token_embeddings` behavior; the rest are similar-by-theme at best but not the same concrete bug.", + "confidence": 0.72, + "canonical_issue_reason": "`issue:45292` is the cleanest representative of the only strong duplicate thread here: it states the generic `resize_token_embeddings` / `output_embeddings` failure, which is the broader form of the more model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:45276` is the best global issue representative because it anchors the biggest coherent duplicate family in the packet. The training-loss double-shift and `lm_head`-tying reports look like separate, smaller subclusters rather than a single better umbrella issue.", + "best_issue_reason": "`issue:45292` is the best global issue choice because it is the most general and concise formulation of the shared embedding-resize bug, making it the best canonical anchor for the accepted duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Similar loading/tokenizer surface area, but different models and failure modes; not clearly one bug." + "reason": "Both are model-loading complaints, but they describe different checkpoints and failure modes, not the same bug." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "Different models and unrelated symptoms; no shared code-path bug." + "reason": "Unrelated models and problems: SigLIP2 docs vs GLM5 loading." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "CUDA device-mapping assert vs embedding/token leakage are unrelated." + "reason": "Different subsystems and symptoms: Qwen auto device mapping crash vs AudioFlamingo3 batched inference leakage." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Different code paths: FSDP2 attention-state rebuild vs multimodal config unwrapping." + "reason": "Both touch model internals, but one is an FSDP attention-state bug and the other is `AutoModelForCausalLM.from_config` unwrapping; not the same issue." }, { "left": "issue:43404", "right": "issue:43454", - "accept": true, - "reason": "Both report missing `lm_head` weight tying causing generation corruption; same underlying bug family." + "accept": false, + "reason": "Both mention `lm_head` tying, but they affect different model families and do not clearly share one concrete code-path bug." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing/LoRA behavior and an apex RMSNorm memory leak are unrelated." + "reason": "Gradient checkpointing/LoRA is unrelated to the T5 apex RMSNorm memory leak." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch version incompatibility and meta-device loading failures are different problems." + "reason": "Torch version compatibility and `meta`-device loading are different problems." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "`init_empty_weights` argument handling and mmap/OOM on Strix Halo are unrelated." + "reason": "`init_empty_weights` argument handling and Strix Halo mmap OOM are unrelated." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both are quantization/memory themed, but the concrete failure modes differ." + "reason": "Both involve quantization, but one is offloading behavior and the other is int4 memory regression causing OOM." }, { "left": "issue:44079", "right": "issue:46829", "accept": false, - "reason": "Different output-mapping bugs in different components; no clear shared fix." + "reason": "Different output/container bugs in different components." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for `GenerationMixin` and Whisper pipeline regression are unrelated." + "reason": "Import path regression vs Whisper pipeline `return_language` behavior are not the same bug." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Both touch quantization, but the affected models and bugs are different." + "reason": "Both are quantization-related, but they concern different formats and failure modes." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Different model-specific bugs: FSDP2 attention state vs CPU-offload rotary embedding." + "reason": "Different model-specific failures: FSDP attention state vs CPU offload incompatibility." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Separate memory issues with different triggers and code paths." + "reason": "Both mention OOM, but one is int4 CUDA reserved memory and the other is mmap on Strix Halo." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "Audio track leakage and device-mismatch during forward are unrelated." + "reason": "Different bug classes: embedding/token leakage vs device mismatch in Gemma4 PLE." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but they are distinct bugs." + "reason": "Both mention gradient checkpointing, but one is ineffective under PEFT LoRA and the other is a mode restriction in eval; not the same bug." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "SwitchTransformers sparse-layer config and Gemma4 attention KeyError are unrelated." + "reason": "Different model/config bugs with no shared concrete failure path." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading breakage and mmap/OOM are different failure modes." + "reason": "Meta-device loading regression is not the same as mmap/OOM behavior." }, { "left": "issue:44863", "right": "issue:46468", "accept": false, - "reason": "Checkpoint loading bug vs broken example; related area, but not the same concrete defect." + "reason": "Same model family, but one is checkpoint loading and the other is an example script; too little evidence they are the same underlying fix." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same `resize_token_embeddings` regression: downstream/output embeddings are not updated after resizing." + "reason": "`issue:45276` explicitly includes the `output_embeddings` failure covered by `issue:45292`; this is the same resize-token-embedding bug, with 45276 just being a broader/more specific report." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Eval-mode checkpointing issue and RLHF NaNs from SDPA fallback are unrelated." + "reason": "Both mention gradient-related instability, but one is checkpointing in eval mode and the other is RLHF NaNs from Qwen3.5 position IDs." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video-input regression and strict config loading are different issues." + "reason": "Different families of regressions: video-input handling vs strict config loading." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "Serialization extra-state handling and mmap/OOM are unrelated." + "reason": "Save/load `extra_state` handling is unrelated to mmap-based OOM." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both involve save/from_pretrained, but permissions propagation and wrong-module loading are different bugs." + "reason": "Both involve save/load of custom modules, but one is permissions propagation and the other is loading the wrong custom module." }, { "left": "issue:42796", "right": "issue:46326", "accept": false, - "reason": "Tokenizer/mistral_common related, but the concrete behaviors differ." + "reason": "Tokenizer BPE-marker handling and `continue_final_message` are related only at a broad tokenizer level, not the same concrete bug." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Flash-attention sinkless-model crash and Gemma4 FSDP2 KeyError are unrelated." + "reason": "Flash-attention sinkless-model crash and Gemma4 FSDP state rebuilding are unrelated." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Unit-test model loading failure and save_pretrained regression are unrelated." + "reason": "Different loading/regression reports for different models and APIs." }, { "left": "issue:45292", "right": "issue:45335", - "accept": true, - "reason": "Same `resize_token_embeddings` family: resized embeddings fail to propagate to other embedding/output paths." - }, - { - "left": "issue:46894", - "right": "issue:46901", - "accept": true, - "reason": "Both describe the same training-loss double-shift pattern; same underlying labels-shift bug family." + "accept": false, + "reason": "Both are about `resize_token_embeddings`, but they target different model implementations (`output_embeddings` vs `decoder.embed_tokens`) and are not clearly the same concrete bug." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Different model-loading failures for different checkpoints; no clear shared bug." + "reason": "Different model-loading reports with different root causes." + }, + { + "left": "issue:46894", + "right": "issue:46901", + "accept": false, + "reason": "Both describe a double-shift training-loss pattern, but they are model-specific reports without enough evidence of one shared concrete code-path fix." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The writeup is a bit too strong for the packet. In addition to the resize_token_embeddings family, the packet also contains another plausible duplicate family around training-loss double-shift (issue:46894 vs issue:46901), and there is a second near-match around lm_head tying (issue:43404 vs issue:43454). Because of that, the claim that this is the *only* coherent duplicate family / that everything else is mostly unrelated overstates the evidence. The soft-edge rejections are mostly conservative, but the 46894/46901 rejection in particular looks too dismissive given the matching bug pattern in the titles." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft edge, issue:45276 vs issue:45292, is supported by the titles and strong similarity, and the rest of the verdicts avoid overclaiming shared bugs. The canonical choice for 45292 is defensible as a concise representative of the resize_token_embeddings/output_embeddings thread." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -2933,8 +2921,8 @@ "issue:43425|issue:43716", "issue:44479|issue:44805", "issue:43906|issue:44530", - "issue:43901|issue:44509", "issue:44589|issue:45084", + "issue:43901|issue:44509", "issue:43994|issue:44079", "issue:36032|issue:43452", "issue:44265|issue:44898", @@ -2951,238 +2939,238 @@ "issue:45479|issue:46897", "issue:44164|issue:46821" ], - "prepared_review_unit_hash": "7f76e112f6966937b20cd3cee1ee9b85669704ff06880cef1378c142b6824bda", + "prepared_review_unit_hash": "475bb83dc03a63ccc0569b9ed97992ab757c01a11a517005917f9cd3900874d4", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 38132, - "estimated_input_tokens": 9533, - "estimated_eval_tokens": 19322 + "serialized_chars": 38136, + "estimated_input_tokens": 9534, + "estimated_eval_tokens": 19324 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers v5 regressions and docs issues. None of the soft pairs look like safe duplicates; they share broad subsystem labels but not the same concrete bug or fix path.", - "confidence": 0.42, - "canonical_issue_reason": "If forced to pick one issue as an umbrella, issue 43611 is the broadest model-loading regression, but it still does not actually subsume the rest of the set.", + "summary": "These items form a loose transformer-v5 regression cluster, but the soft pairs mostly describe distinct bugs, model families, or code paths rather than true duplicates. No PRs were provided.", + "confidence": 0.72, + "canonical_issue_reason": "issue:43611 is the best representative because it describes a concrete, broadly relevant v5 loading regression and is more general than the highly model-specific reports around it.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43611 is the least specific and most central to the repeated v5 model-loading complaints, so it is the best weak canonical representative for the cluster.", + "best_issue_reason": "issue:43611 is the clearest single issue to anchor the cluster; it is specific enough to be actionable yet broad enough to stand in for related v5 model-loading breakages.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "One is a custom-model initialization regression; the other is a Trainer resume/max_steps accounting bug. Different code paths and fixes." + "reason": "Different problems: custom-model initialization vs trainer checkpoint resumption/max_steps calculation." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook custom-model initialization failures are not the same as save/from_pretrained mishandling extra_state." + "reason": "Notebook custom-model initialization and extra_state save/load handling are separate failure modes." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE is a different failure mode from meta-device loading breakage." + "reason": "One is an autocast dtype mismatch for a specific model; the other is meta-device loading breakage." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to the Strix Halo mmap/OOM issue." + "reason": "Float8 storage lookup and mmap OOM on Strix Halo are unrelated runtime issues." }, { "left": "issue:46315", "right": "issue:46491", "accept": false, - "reason": "Tokenizer performance regression is not the same bug as a CodeLlama encode-decode space-loss regression." + "reason": "Tokenizer conversion slowdown is not the same as CodeLlama space-preservation regression." }, { "left": "issue:45812", "right": "issue:46315", "accept": false, - "reason": "Wrong Granite token IDs and slow convert_tokens_to_ids are distinct tokenizer problems." + "reason": "Wrong token IDs for Granite is a correctness regression, not the convert_tokens_to_ids performance bug." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch 2.10 incompatibility is too broad and not the same concrete autocast dtype mismatch bug." + "reason": "Torch version incompatibility is too broad and does not match the specific autocast mismatch bug." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline return_language breakage is unrelated to a general tokenization-version change report." + "reason": "Whisper pipeline return_language behavior and generic tokenization changes are different regressions." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "PagedAttentionCache linear_attention crash and template-node compilation error are different failures." + "reason": "Qwen3.5 PagedAttentionCache crash and template-node compilation failure are different code paths." }, { "left": "issue:39404", "right": "issue:46665", "accept": false, - "reason": "Pipeline return_language regression is not the same as inconsistent from_pretrained behavior inside Python classes." + "reason": "Pipeline return_language regression is unrelated to from_pretrained behavior inside Python classes." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both involve model loading, but one is about base_model_prefix handling and the other about notebook-defined custom models; not the same bug." + "reason": "Both involve loading/custom models, but the concrete bugs and triggers are different." }, { "left": "issue:43425", "right": "issue:46869", "accept": false, - "reason": "Torch 2.10 incompatibility is not the same as a YOLO multi-GPU device-map failure." + "reason": "A torch-version incompatibility is not the same as YOLO multi-GPU device-map failure." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention torch-version breakage is unrelated to Gemma4 PLE device mismatch." + "reason": "Compiled flex_attention on new torch versions and Gemma4 device placement are unrelated." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "A general torch compatibility complaint does not match a specific device-mapping forward-pass bug." + "reason": "Generic torch incompatibility does not identify the same device-map bug." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Torch version incompatibility is not the same as an image-preprocessor/model dtype mismatch." + "reason": "Torch compatibility and image preprocessor/model dtype mismatch are distinct." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression for Qwen models and a mask-shape IndexError are different concrete bugs." + "reason": "Video-input regression for Qwen models and mask-shape IndexError are different symptoms and likely different fixes." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "An isolated reproduction of another issue is not evidence of the same PagedAttentionCache crash." + "reason": "Isolated reproduction of one prior issue is not the same bug as the PagedAttentionCache linear_attention crash." }, { - "left": "issue:43901", - "right": "issue:44509", + "left": "issue:44589", + "right": "issue:45084", "accept": false, - "reason": "Both are docs-related, but they target different deprecated/removed pipeline task mentions." + "reason": "Float8 storage-object errors are unrelated to template compilation errors." }, { - "left": "issue:44589", - "right": "issue:45084", + "left": "issue:43901", + "right": "issue:44509", "accept": false, - "reason": "Float8 storage lookup failure and template compilation failure are unrelated." + "reason": "Docs mentioning deprecated pipeline tasks is separate from behavior/docs mismatch for return_all_scores." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Incorrect SigLIP2 outputs are not the same issue as ModelOutput key assignment for None-valued keys." + "reason": "Siglip2 wrong outputs and ModelOutput key assignment are different underlying problems." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "T5 tokenizer method conflict and gguf_file failures in from_pretrained are separate loading paths." + "reason": "T5 tokenizer special-tokens conflict is unrelated to gguf_file loading failures." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export/torch_compilable_check failure is unrelated to Perceiver non-default-resolution classification." + "reason": "torch.export exportability with torch_compilable_check and Perceiver interpolation failure are not the same bug." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline regression and MXFP4 quantization fallback are different model families and bugs." + "reason": "Whisper return_language regression and MXFP4 quantization fallback are unrelated." }, { "left": "issue:46050", "right": "issue:46326", "accept": false, - "reason": "Quantized save/load breakage is unrelated to mistral_common continue_final_message behavior." + "reason": "Quantized save/load breakage and mistral_common continue_final_message are different features and code paths." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "extra_state persistence and composite-config text_config unwrapping are distinct serialization/config bugs." + "reason": "extra_state save/load handling is unrelated to AutoModelForCausalLM config unwrapping." }, { "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "An isolated reproduction of a prior issue is not the same as a tokenizer continue_final_message bug." + "reason": "A reproduction issue for an older bug does not match the mistral_common tokenizer behavior change." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM prediction errors are not the same as tokenizer loading failures for est-roberta." + "reason": "CamemBERT masked-LM outputs and tokenizer-loading failure for EMBEDDIA/est-roberta are separate." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizers, but the affected models and failing code paths differ; not a safe duplicate." + "reason": "Both are tokenizer-related, but one is a loading failure and the other is a regex patch crash for Mistral." }, { "left": "issue:45823", "right": "issue:46899", "accept": false, - "reason": "Gemma4 device mismatch with auto device_map is different from unsafe dtype casting in multimodal embedders." + "reason": "Gemma4 device-map mismatch and Gemma4 multimodal 4-bit dtype casting are different failures." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both are Qwen3-VL loading problems, but one is a dev0 model-loading regression and the other is a weight-shape mismatch; not clearly the same fix." + "reason": "Both concern Qwen3-VL loading, but the reported failures are different enough that they should not be merged." }, { "left": "issue:45479", "right": "issue:46894", "accept": false, - "reason": "Zero-loss sequence-classification bug and ASR double-shift training-loss bug are different training targets and failure modes." + "reason": "Zero-loss classification bug and training-loss double-shift bug are different model/task failures." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Qwen3 MoE training memory usage and Strix Halo mmap OOM are not the same underlying issue." + "reason": "Qwen3 MoE training memory usage and Strix Halo mmap OOM are unrelated." }, { "left": "issue:45479", "right": "issue:46897", "accept": false, - "reason": "Different model families and different loss-path symptoms; too weak to merge." + "reason": "Single-label classification zero-loss regression is not the same as Florence2 double-shift training bug." }, { "left": "issue:44164", "right": "issue:46821", "accept": false, - "reason": "save/from_pretrained extra_state handling is unrelated to a TimesFM window_size AttributeError." + "reason": "extra_state persistence and TimesFM window_size AttributeError are unrelated." } ] }, @@ -3262,8 +3250,8 @@ "issue:43611|issue:43644", "issue:44462|issue:46489", "issue:42796|issue:43404", - "issue:46285|issue:46402", "issue:36032|issue:42915", + "issue:46285|issue:46402", "issue:39404|issue:44671", "issue:43716|issue:45084", "issue:43643|issue:44530", @@ -3274,8 +3262,8 @@ "issue:44451|issue:44987", "issue:43644|issue:43646", "issue:34567|issue:40444", - "issue:40444|issue:46326", "issue:43541|issue:45084", + "issue:40444|issue:46326", "issue:42175|issue:46178", "issue:43582|issue:43632", "issue:44206|issue:44220", @@ -3288,14 +3276,14 @@ "issue:45084|issue:45507", "issue:43746|issue:46326" ], - "prepared_review_unit_hash": "5d5b1452e41f3b622ce3860e3210f8aaa5aaad483e7466b25fa41441e9bb8e01", + "prepared_review_unit_hash": "b5794be8d42c578a6bef2b0b4184f3c3c73aa8a9d3da29c8c55cdbfa6305787e", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, @@ -3310,208 +3298,205 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most candidate pairs are related only at a broad subsystem or version-regression level, not the same concrete bug. The only strong duplicate-like pair is the GGUF/from_pretrained loading regression.", - "confidence": 0.66, - "canonical_issue_reason": "issue:43452 is the broadest and most general report of the GGUF/from_pretrained loading break, and issue:43482 looks like a model-specific manifestation of the same failure.", + "summary": "This cluster is a loose grab-bag of unrelated Transformers regressions and model/tokenizer edge cases, with only weak thematic overlap around v5 loading behavior. None of the soft pairs are strong enough duplicates to merge.", + "confidence": 0.91, + "canonical_issue_reason": "No single issue is a safe canonical for the cluster: the items span distinct bugs across tokenizers, model initialization/loading, audio, quantization, and runtime/device behavior.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43452 is the best representative because it describes the underlying loading regression generically rather than only for one GGUF model.", + "best_issue_reason": "issue:43646 is the broadest representative of the cluster\u2019s recurring v5 regression theme, but it is still too generic to absorb the others as duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "Different bugs: zero3/BertModel indexing versus flex_attention compilation on torch>=2.9." + "reason": "Different failures and code paths: Deepspeed/ZERO3 BertModel indexing vs torch>=2.9 flex_attention compilation." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Both are loading failures, but for different models and likely different tokenizer/model-resolution paths." + "reason": "Both are loading regressions, but they concern different models and likely different tokenizer/model paths." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Different tokenizer failures: one cannot load a BERT model, the other is a Mistral regex patch crash." + "reason": "BERT model loading failure is not the same bug as the Mistral tokenizer regex patch crash." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffer initialization." + "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffer initialization corruption." }, { "left": "issue:44462", "right": "issue:46489", "accept": false, - "reason": "Related to tokenizer loading, but one is repository tokenizer.json selection and the other is a DeepSeek-Coder-specific output regression." + "reason": "Both involve tokenizer loading/output, but the titles point to different model-specific bugs and different symptoms." }, { "left": "issue:42796", "right": "issue:43404", "accept": false, - "reason": "Mistral tokenizer BPE-marker handling and Mistral3 lm_head tying are unrelated code paths." + "reason": "Tokenizer BPE-marker handling and lm_head weight tying are separate Mistral3 issues." }, { - "left": "issue:46285", - "right": "issue:46402", + "left": "issue:36032", + "right": "issue:42915", "accept": false, - "reason": "Different model classes and failures: CLIP structure mismatch versus Sam3 text encoder weights not loading." + "reason": "T5 tokenizer load error and Qwen3Moe FineGrainedFP8 failure are unrelated." }, { - "left": "issue:36032", - "right": "issue:42915", + "left": "issue:46285", + "right": "issue:46402", "accept": false, - "reason": "Tokenizer add_special_tokens conflict is unrelated to Qwen3MoE fine-grained FP8 loading." + "reason": "CLIP structure inconsistency is unrelated to Sam3 text encoder weights not loading." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline return_language and CamemBERT masked-LM prediction issues are unrelated." + "reason": "Whisper pipeline return_language regression and CamemBERT masked-LM prediction regression are unrelated." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Mistral image dtype mismatch is unrelated to the template-node compilation error." + "reason": "Image preprocessor/model dtype mismatch is a different bug from the template-node compile error." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "AutoConfig trust_remote_code missing fields is unrelated to PagedAttentionCache linear_attention failures." + "reason": "trust_remote_code config field loss and Qwen3.5 paged-attention invalid group type are unrelated." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "Both touch device placement, but XLNet CPU arange and Qwen3.5MoE CPU offload are different bugs." + "reason": "Both mention device placement, but they affect different models and different code paths." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Tokenizer mapping-name failure and flex_attention compile failure are unrelated." + "reason": "Tokenizer mapping-name validation and flex_attention compilation are unrelated." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Perplexity anomaly and runtime special-token post-processor update bug are different problems." + "reason": "Perplexity regression on a benchmark and runtime special-token post-processor updates are not the same bug." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Qwen3MoE FP8 config failure is unrelated to PagedAttentionCache linear_attention crashes." + "reason": "Same broad Qwen family, but FP8 config failure and paged-attention cache crash are different issues." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Both are model loading failures, but with different models and no clear shared code-path." + "reason": "Both are latest-version loading failures, but for different repositories and likely different root causes." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Non-persistent buffer junk and custom model initialization are adjacent v5 regressions, not the same concrete bug." + "reason": "Related to v5 initialization, but the symptoms differ: buffer corruption vs custom model initialization breakage." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "TrainerState token counting and multimodal iterable dataset failures are unrelated." + "reason": "TrainerState token accounting and Qwen2.5-VL iterable dataset failures are unrelated." }, { - "left": "issue:40444", - "right": "issue:46326", + "left": "issue:43541", + "right": "issue:45084", "accept": false, - "reason": "Multiple-image dataset finetuning and Mistral-common continue_final_message handling are unrelated." + "reason": "Torch dynamo grouped_mm tracing failure is unrelated to the non-template-node compile error." }, { - "left": "issue:43541", - "right": "issue:45084", + "left": "issue:40444", + "right": "issue:46326", "accept": false, - "reason": "Dynamo tracing grouped_mm failure is unrelated to the non-template-node compilation error." + "reason": "Multimodal training with multiple images and Whisper tokenizer continue_final_message are different bugs." }, { "left": "issue:42175", "right": "issue:46178", "accept": false, - "reason": "Backend packaging omission and BACKENDS_MAPPING tensorflow_text error are different backend issues." + "reason": "Backend packaging/installation issue and tensorflow_text backend mapping error are related to dependencies, not the same defect." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "AppleSilicon warmup TypeError and _is_hf_initialized flag breakage are unrelated." + "reason": "AppleSilicon allocator warmup TypeError and _is_hf_initialized regressions are separate initialization problems." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Both are audio-feature related, but one is a center-arg regression in LasrFeatureExtractor and the other is a separate _torch_extract_fbank_features issue." + "reason": "Both are audio-feature related, but the titles do not establish the same root cause with enough confidence." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both involve Whisper post-processing, but batch_decode skip_special_tokens and empty-transcription generation after align_special_tokens are different bugs." + "reason": "Whisper batch_decode skip_special_tokens behavior and empty-transcription generation after align_special_tokens are different symptoms." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "T5 tokenizer add_special_tokens conflict is unrelated to custom model initialization in notebooks." + "reason": "Tokenizer class-method conflict and custom model initialization in notebooks are unrelated." }, { "left": "issue:43452", "right": "issue:43482", - "accept": true, - "reason": "Both point to the same GGUF/from_pretrained loading regression; 43482 looks like a specific model instance of the generic failure in 43452." + "accept": false, + "reason": "GGUF from_pretrained breakage and Qwen2.5-GGUF loading failure are both GGUF-related but not clearly the same bug." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing all_tied_weights_keys and failure to load cjvt/sleng-bert are unrelated." + "reason": "Molmo tied-weights attribute error and BERT model loading failure are unrelated." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Different model-specific tokenizer loading failures; no evidence of the same underlying bug." + "reason": "Both are transformer v5 tokenizer-load failures, but they affect different models and likely different tokenizer issues." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "GPU/CPU offload dequantization loading and GGUF v5 loading failure are different issues." + "reason": "Dequantized model offload failure is unrelated to Qwen2.5-GGUF loading failure." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Template compilation error and GraniteMoEHybrid invalid method call are unrelated." + "reason": "Template compilation failure and GraniteMoEHybrid invalid method calls are different bugs." }, { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "PEFT local-checkpoint loading and Mistral-common continue_final_message are unrelated." + "reason": "PEFT local-checkpoint loading for GraniteSpeech is unrelated to Whisper continue_final_message behavior." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The judgments are conservative and grounded in the packet titles. The only accepted duplicate-like pair, issue:43452 vs issue:43482, is a reasonable generic-vs-model-specific GGUF loading match. The rest of the rejections avoid overclaiming and stay at the level supported by the available evidence. " - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -3581,8 +3566,8 @@ "issue:43262|issue:43377", "issue:43646|issue:44534", "issue:43645|issue:46661", - "issue:44849|issue:45325", "issue:44403|issue:45081", + "issue:44849|issue:45325", "issue:42915|issue:45084", "issue:43377|issue:43819", "issue:43883|issue:44991", @@ -3605,228 +3590,228 @@ "issue:36032|issue:41720", "issue:43632|issue:46821", "issue:44991|issue:46315", - "issue:45216|issue:45357", "issue:43531|issue:44877", + "issue:45216|issue:45357", "issue:40444|issue:45325" ], - "prepared_review_unit_hash": "83721c102e988734d77a8c7bff80339ebe0de76afcc82b864e3aa477778f261a", + "prepared_review_unit_hash": "d226bb892b17e40552473eca605b844ec34037739f33a371a40030df4932cf9e", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36934, - "estimated_input_tokens": 9234, - "estimated_eval_tokens": 18724 + "serialized_chars": 36937, + "estimated_input_tokens": 9235, + "estimated_eval_tokens": 18726 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Possible thematic subclusters rather than confirmed duplicate clusters. The strongest theme is a broad Transformers v5 custom-model initialization regression, best anchored by issue:43646; the notebook-specific issue:43645 looks like a narrower manifestation. A separate Qwen VL/video-positioning theme appears, but the pair evidence is still only soft similarity and not duplicate-level proof.", - "confidence": 0.29, - "canonical_issue_reason": "issue:43646 is the broader custom-model initialization report, while issue:43645 is the narrower Jupyter-notebook-specific manifestation, so 43646 is the better canonical anchor.", + "summary": "Mostly a loose topical cluster of unrelated Transformer regressions. I\u2019d keep only the Qwen3.5 save_pretrained pair (45216\u219445357) as a clear duplicate; the tokenizer AddedToken and Qwen VL rope-index pairs look too weakly supported to merge confidently.", + "confidence": 0.77, + "canonical_issue_reason": "Issue 45216 is the clearest canonical representative: it states the Qwen3.5 save_pretrained regression directly, and 45357 reads like a narrower follow-up report of the same incorrect visual-encoder-keys problem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the clearest broad representative of the custom-model initialization theme and is better suited as the cluster anchor than the notebook-specific follow-up.", + "best_issue_reason": "45216 is the best standalone issue anchor because it describes a concrete regression with a direct duplicate/follow-up report, making the bug scope easiest to interpret.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Different subsystems and failure modes: audio processor sampling-rate default vs MIMI batching/padding-mask behavior." + "reason": "Both are audio-related, but they describe different failures: chat-template sampling-rate defaulting vs batched padding-mask behavior in MIMI." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Both are v5 regressions, but they affect different code paths: custom model initialization vs non-persistent buffer serialization." + "reason": "Custom model initialization breakage and non-persistent buffer corruption are different code paths and different bugs." }, { "left": "issue:43645", "right": "issue:46661", "accept": false, - "reason": "Only broad thematic overlap around model loading/init; notebook-specific custom model initialization and remote-code model_class.config_class errors are not concrete evidence of the same defect." + "reason": "Both involve model loading, but the failure modes differ: notebook custom-model initialization vs missing config_class on remote-code models." }, { - "left": "issue:44849", - "right": "issue:45325", + "left": "issue:44403", + "right": "issue:45081", "accept": false, - "reason": "Both are Qwen-related, but they describe different bugs: output_hidden_states handling vs still-image temporal position scaling in rope indexing." + "reason": "Unnecessary load-time noise and a Mistral tokenizer regex patch crash are unrelated symptoms with different fixes." }, { - "left": "issue:44403", - "right": "issue:45081", + "left": "issue:44849", + "right": "issue:45325", "accept": false, - "reason": "Unnecessary loading noise and a tokenizer regex crash are unrelated failure modes." + "reason": "These are different Qwen regressions: output_hidden_states handling in Qwen3.5 versus rope-index temporal-position scaling in Qwen2.5-VL." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Jinja/template compilation error are different code paths and causes." + "reason": "FineGrainedFP8Config failure and non-template-node compilation are distinct issues." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "Different models and mechanisms: MIMI padding-mask support vs DAC.from_latents / STE mismatch." + "reason": "Both are multimodal/audio-model bugs, but one is missing padding-mask support and the other is DAC from_latents disagreeing with forward due to missing STE." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Missing all_tied_weights_keys on Molmo and a tokenizer loading failure for another model are unrelated." + "reason": "Missing all_tied_weights_keys in Molmo is unrelated to the tokenizer-loading regression for EMBEDDIA/est-roberta." }, { "left": "issue:43792", "right": "issue:46869", "accept": false, - "reason": "Whisper runtime failure and YOLO multi-GPU device-map failure are not the same bug." + "reason": "Whisper load failure and YOLO multi-GPU auto device-map failure are separate model-specific problems." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Both mention tied weights, but one is a model-specific tying bug and the other is a serialization/device-dependent save issue." + "reason": "They both mention tied weights, but one is a model-specific tying bug and the other is a serialization inconsistency across device placement." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward and StaticCache.get_seq_length type mismatch are separate bugs." + "reason": "Compiled-forward chunked generation inconsistency is not the same as StaticCache.get_seq_length returning the wrong type." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Different problems: incorrect SigLIP2 outputs vs processor/model image-size mismatch." + "reason": "Siglip2 AutoModel/pipeline nonsense outputs and OmDet-Turbo processor size mismatch are different vision bugs." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Different affected models and root causes; loading failure vs Mistral tokenizer regex patch crash." + "reason": "A model-specific tokenizer load regression and a Mistral tokenizer regex patch crash are not the same underlying bug." }, { "left": "issue:28530", "right": "issue:34567", "accept": false, - "reason": "Early stopping metric selection and token-count tracking are unrelated training issues." + "reason": "Early stopping metric selection and num_input_tokens_seen not updating are unrelated Trainer issues." }, { "left": "issue:44008", "right": "issue:46899", "accept": false, - "reason": "Gemma3n variable-name collision and Gemma4 quantization dtype cast are different multimodal bugs." + "reason": "Gemma3n variable collision in forward and Gemma4 multimodal dtype-cast corruption are different bugs." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "FineGrainedFP8Config failure and unexpected _is_hf_initialized argument error are separate initialization/config issues." + "reason": "FineGrainedFP8Config failure and init_empty_weights/_is_hf_initialized argument breakage are separate initialization problems." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "ScandiBERT loading failure is not the same as the Mistral tokenizer backend_tokenizer crash." + "reason": "ScandiBERT loading failure is not the same as the Mistral regex tokenizer crash." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "T5 tokenizer add_special_tokens conflict and MobileLLM key error are unrelated tokenizer/model-loading bugs." + "reason": "T5 add_special_tokens method conflict and MobileLLM key error are different tokenizer/model-loading issues." }, { "left": "issue:44062", - "right": "issue:44987", + "right": "issue:44991", "accept": false, - "reason": "AddedToken special-argument duplication and a model loading failure are different tokenizer issues." + "reason": "The evidence is too thin: one side is a generic tokenizer-loading regression and the other is a specific AddedToken duplicate-special TypeError." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export / torch_compilable_check failure is unrelated to OmDet-Turbo processor size mismatch." + "reason": "torch.export / torch_compilable_check failure is unrelated to the OmDet-Turbo processor size regression." }, { "left": "issue:43089", "right": "issue:45987", "accept": false, - "reason": "Generation performance overhead and StaticCache.get_seq_length type mismatch are not the same bug." + "reason": "Generation overhead from GPU syncs/dispatch is a performance issue, not the seq_length return-type bug." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Llava Next loading segfault and Qwen3OmniMoe video-input ValueError are different multimodal failures." + "reason": "Llava-Next loading segfault and Qwen3OmniMoe video-input ValueError are unrelated multimodal failures." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "The packet only shows thematic Qwen VL similarity; video-input regression and still-image rope scaling are distinct symptom paths without concrete shared defect evidence." + "reason": "This mixes a broad video-input regression with a still-image rope-index bug; the shared Qwen VL theme is not enough to treat them as the same underlying issue." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "FP8 config failure and missing Float8 storage type are different compatibility issues." + "reason": "FineGrainedFP8Config failure and missing Float8 storage object are different quantization/runtime errors." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "T5 tokenizer construction error and Qwen3 auto device mapping assert are unrelated." + "reason": "Tokenizer method-name collision and Qwen3 auto device-map cuda assert are unrelated." }, { "left": "issue:43632", "right": "issue:46821", "accept": false, - "reason": "_is_hf_initialized flag breakage and TimesFM window_size attribute error are unrelated v5 regressions." + "reason": "_is_hf_initialized breakage and TimesFM window_size AttributeError are distinct v5 regressions." }, { "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "Tokenizer loading failure for one model and slow convert_tokens_to_ids regression are different issues." - }, - { - "left": "issue:45216", - "right": "issue:45357", - "accept": false, - "reason": "Related Qwen3.5 save_pretrained area, but checkpoint correctness and visual-encoder key serialization are not the same concrete bug." + "reason": "Both are tokenizer-related, but one is a load failure and the other is a performance regression in convert_tokens_to_ids." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Qwen3-MoE sliding_window bug and granite_speech strict-config loading failure are unrelated." + "reason": "Qwen3-MoE sliding_window behavior and granite_speech strict-config loading are unrelated issues." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "These are the same Qwen3.5 save_pretrained regression; the later report just narrows the symptom to incorrect visual encoder keys." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Multi-image IterableDataset finetuning failure is unrelated to Qwen2.5-VL temporal position scaling." + "reason": "Multiple-images-per-prompt finetuning failure and still-image temporal-position scaling are different Qwen2.5-VL bugs." } ] }, "evaluator_result": { "accept": false, - "feedback": "The write-up is not conservative enough. The packet only supports broad thematic similarity, not duplicate-level evidence, for the two accepted pairs. In particular, issue:43645 vs issue:46661 and issue:44479 vs issue:45325 are both marked accept despite the packet giving only soft similarity and no concrete shared defect evidence; they should stay rejected. Also, the canonical anchor choice is a bit overstated: issue:43646 is the broader custom-model initialization report, while issue:43645 is the narrower notebook-specific one. Overall, the summary should be downgraded to \"possible thematic subclusters\" rather than confirmed duplicate clusters." + "feedback": "The overall framing is conservative and grounded, but two of the accepted soft-edge matches are too assertive for the evidence in the packet. In particular, 44062\u219444991 is only supported by a generic tokenizer-loading regression on one side and a specific AddedToken TypeError on the other, and 44479\u219445325 mixes a broad video-input regression with a still-image rope-index bug. Those should stay tentative unless the issue bodies clearly line up. The 45216\u219445357 duplicate looks fine, but because the cluster summary relies on the other two accepted pairs being duplicates, I\u2019d reject for overstating evidence." }, "evaluator_used": true, "retried": true, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -3912,236 +3897,233 @@ "issue:35707|issue:40444", "issue:43482|issue:43525", "issue:43421|issue:44568", - "issue:43482|issue:45538", "issue:43632|issue:44534", + "issue:43482|issue:45538", "issue:43531|issue:44291", "issue:41553|issue:43582", "issue:42175|issue:45542", "issue:41720|issue:45081", "issue:43317|issue:43856", - "issue:44462|issue:45812", "issue:44079|issue:44534", - "issue:44509|issue:46909", - "issue:43377|issue:45491" + "issue:44462|issue:45812", + "issue:43377|issue:45491", + "issue:44509|issue:46909" ], - "prepared_review_unit_hash": "eb249ed966b54f4e5c5475ecc5819067765e2883907ef7c9561fd7c2892b7d00", + "prepared_review_unit_hash": "3fb937f89ee49f87516083b647d53353285cf2c49de871aae59ca596513d4bb7", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36631, - "estimated_input_tokens": 9158, - "estimated_eval_tokens": 18572 + "serialized_chars": 36621, + "estimated_input_tokens": 9156, + "estimated_eval_tokens": 18568 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated bugs; only the AutoTokenizer/tokenizer.json regression pair looks like a likely duplicate. I\u2019d use the broad tokenizer issue as the best anchor and keep the rest separate.", - "confidence": 0.41, - "canonical_issue_reason": "issue:44462 is the broadest, most general tokenizer regression here and best matches the one plausible duplicate pair in the cluster.", + "summary": "Mostly unrelated issues; the soft pairs only share broad themes like v5 regressions, tokenizers, or multimodal training, but not the same underlying bug/change. I would not merge any of the suggested pairs.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out because this is a heterogeneous set; the soft pairs do not point to one shared defect. If forced to pick an anchor, issue:44534 is the most generic core-regression report, but it is not a duplicate hub.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44462 is the best global issue anchor because it is model-agnostic and captures the shared AutoTokenizer/tokenizer.json regression more cleanly than the model-specific reports.", + "best_issue_reason": "No issue is a strong global best duplicate representative here. The cluster is too mixed, and the candidate pairs are only loosely related by subsystem, not by exact bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "Both are Transformers v5 regressions, but one is about base_model_prefix loading and the other about non-persistent buffers; different code paths." + "reason": "Both are Transformers v5 regressions, but they describe different failures: base_model_prefix loading vs junk-filled non-persistent buffers." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch version incompatibility and noisy load warnings are unrelated problems." + "reason": "Unrelated problems: Torch 2.10 compatibility versus noisy logging during model loading." }, { "left": "issue:43716", "right": "issue:46899", "accept": false, - "reason": "Both mention multimodal dtype issues, but one is an image-preprocessor mismatch and the other is a 4-bit audio/vision quantization cast bug." + "reason": "Both are multimodal dtype issues, but one is image preprocessor/model mismatch and the other is 4-bit unsafe casting in Gemma4 embedders." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "PaddleOCR-VL loading failure and Qwen3-MoE sliding_window behavior are unrelated." + "reason": "Different models and failure modes: PaddleOCR-VL load failure vs Qwen3-MoE sliding_window behavior." }, { "left": "issue:45812", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but Granite wrong IDs and CodeLlama leading-space loss are different symptoms and likely different fixes." + "reason": "Both concern tokenization, but one is Granite token IDs and the other is a CodeLlama leading-space round-trip regression." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Fbank extraction and compiled flex_attention are unrelated subsystems." + "reason": "Audio feature extraction bug versus compiled flex_attention failure on newer Torch versions." }, { "left": "issue:42796", "right": "issue:43746", "accept": false, - "reason": "Tokenizer BPE marker handling and GraniteSpeech PEFT loading are not the same bug." + "reason": "Tokenizer BPE-marker handling and GraniteSpeech PEFT checkpoint loading are separate bugs." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing in eval mode and a missing supports_gradient_checkpointing flag are distinct issues." + "reason": "Gradient checkpointing in eval mode is not the same as a missing supports_gradient_checkpointing flag on a trust_remote_code model." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Generic loading noise and AutoTokenizer ignoring tokenizer.json are different bugs." + "reason": "Generic loading noise and a specific AutoTokenizer regression for EMBEDDIA/est-roberta are unrelated." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "_is_hf_initialized handling and mmap OOM on Strix Halo are unrelated." + "reason": "_is_hf_initialized regression and Strix Halo mmap OOM are different code paths and symptoms." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both involve Qwen VL training, but multi-GPU training and iterable-dataset multi-image finetuning failures are different problems." + "reason": "Both involve Qwen2-VL-family training, but one is multi-GPU training and the other is IterableDataset/multiple-images finetuning." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "High perplexity on gpt-oss and Whisper batch_decode skip_special_tokens are unrelated." + "reason": "Perplexity regression on gpt-oss-20b is unrelated to Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper load failure and Float8 storage lookup error do not describe the same bug." + "reason": "Whisper model load/runtime failure and missing Float8 storage are not the same bug." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized and Float8 storage resolution are different failure modes." + "reason": "init_empty_weights/_is_hf_initialized TypeError is distinct from the Float8 storage-object error." }, { "left": "issue:42796", "right": "issue:46710", "accept": false, - "reason": "Tokenizer BPE-marker handling and DeepSeek accuracy regression are unrelated." + "reason": "Tokenizer BPE-marker bug and DeepSeek accuracy regression are different issues, despite both being model-specific regressions." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment and Gemma4 FSDP2 shared_kv_states rebuilding are unrelated." + "reason": "ModelOutput key assignment and Gemma4 shared_kv_states KeyError are unrelated implementation bugs." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL iterable-dataset finetuning." + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to multi-image iterable finetuning failures." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "Qwen2.5-GGUF loading and Llama4Config pad_token_id are different model/config issues." + "reason": "Both are v5 loading failures, but one is Qwen2.5-GGUF compatibility and the other is missing pad_token_id on Llama4Config." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Runtime special-token post-processor updates and add_special_tokens BOS/EOS behavior are related in theme but not clearly the same concrete bug." + "reason": "Runtime special-token post-processor updates and mdeberta-v3-base add_special_tokens behavior are related only at a very broad tokenizer level, not the same concrete bug." }, { - "left": "issue:43482", - "right": "issue:45538", + "left": "issue:43632", + "right": "issue:44534", "accept": false, - "reason": "GGUF loading regression and CLIPTokenizer model_max_length are unrelated." + "reason": "Different Transformers v5 regressions: _is_hf_initialized flag handling versus non-persistent buffer initialization." }, { - "left": "issue:43632", - "right": "issue:44534", + "left": "issue:43482", + "right": "issue:45538", "accept": false, - "reason": "_is_hf_initialized and non-persistent buffer junk are different Transformers v5 regressions." + "reason": "GGUF loading failure and CLIPTokenizer model_max_length overflow are unrelated." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding_window and unexpected _is_hf_initialized arg are not the same issue." + "reason": "Qwen3-MoE sliding_window behavior and init_empty_weights argument handling are distinct issues." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "Bad AutoTokenizer error messaging and Apple Silicon TypeError in caching_allocator_warmup are unrelated." + "reason": "Bad AutoTokenizer error messaging for Voxtral is unrelated to the Apple Silicon TypeError in caching_allocator_warmup." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both involve TensorFlow backend packaging, but one is missing TF in the torch extra and the other is an undefined backend from tensorboard-only installs; not clearly the same bug." + "reason": "Both mention TensorFlow packaging, but one is missing TF in the torch extra and the other is a tensorboard-only undefined backend error." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device mapping crash and Mistral regex patch crash are unrelated." + "reason": "Qwen3 auto device mapping cudaErrorAssert and Mistral tokenizer regex patch crashes are different problems." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "GPU+CPU offload loading failure and Qwen3 MoE memory usage are different problems." - }, - { - "left": "issue:44462", - "right": "issue:45812", - "accept": true, - "reason": "Both point to the same AutoTokenizer/tokenizer.json regression: repository tokenizer config is being ignored, causing wrong token IDs on Granite models." + "reason": "GPU+CPU offload loading failure and Qwen3 MoE training memory inefficiency are not the same bug." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment and v5 non-persistent buffer initialization are unrelated." + "reason": "ModelOutput key assignment and junk-filled non-persistent buffers are different core behaviors." }, { - "left": "issue:44509", - "right": "issue:46909", + "left": "issue:44462", + "right": "issue:45812", "accept": false, - "reason": "Removed pipeline docs and a broken DIA documentation example are both docs issues but not the same change." + "reason": "Ignoring tokenizer.json and Granite token-ID regression are both tokenizer issues but not the same code-path problem." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "MIMI encoder batching/padding-mask bug and Gemma3 NaN embeddings are different model-path failures." + "reason": "MIMI encoder padding-mask mismatch and Gemma3 NaN embeddings from sliding-window attention are separate model bugs." + }, + { + "left": "issue:44509", + "right": "issue:46909", + "accept": false, + "reason": "Removed pipeline docs lingering in v5 and a broken DIA docs example are distinct documentation issues." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is sufficiently grounded in the packet and stays mostly conservative: it treats the cluster as largely unrelated and only elevates one tokenizer-related pair as a plausible duplicate. The negative soft-edge verdicts are appropriately cautious. The 44462/45812 match is still somewhat inferential from titles alone, but not overstated enough to reject." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -4212,8 +4194,8 @@ "issue:43818|issue:46402", "issue:44829|issue:45479", "issue:43644|issue:44079", - "issue:43818|issue:45357", "issue:42175|issue:43296", + "issue:43818|issue:45357", "issue:43493|issue:43994", "issue:44265|issue:45072", "issue:40444|issue:45381", @@ -4239,20 +4221,20 @@ "issue:43881|issue:44488", "issue:41720|issue:45823" ], - "prepared_review_unit_hash": "f34143294708becc6aa303538e54b367ffe98d604d67aaebc0dbd0452d4f1439", + "prepared_review_unit_hash": "f0504f34482e2e1c875c6358df39c3bda87d24430426a870c5c8289ce1a84d0a", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36521, + "serialized_chars": 36522, "estimated_input_tokens": 9131, "estimated_eval_tokens": 18518 }, @@ -4261,204 +4243,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only 2 soft pairs look like true duplicates: one SigLIP2 implementation/regression pair and one Qwen2.5-VL video/position-id regression pair. The rest are broad topical overlaps but describe different failures, models, or code paths.", - "confidence": 0.81, - "canonical_issue_reason": "issue:45325 is the clearest root-cause report among the accepted duplicates: it pinpoints the wrong `get_rope_index` scaling for Qwen2.5-VL visual temporal positions, which subsumes the symptom-only report in issue:45381.", + "summary": "Most pairs are only loosely related by model family or regression theme, not the same underlying bug. The only likely duplicate-looking pair is the SigLIP2 implementation issue; the rest should stay separate.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43493 is the strongest canonical issue candidate because it describes the underlying SigLIP2 implementation discrepancy directly and broadly, while issue:43994 reads more like a user-facing symptom of the same problem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45325 is the strongest representative issue overall because it states the concrete bug mechanism and has the broadest explanatory value for the accepted Qwen2.5-VL duplicate pair.", + "best_issue_reason": "issue:43493 is the best overall issue to anchor this set: it is specific enough to identify the root cause, but broad enough to cover the downstream bad outputs reported elsewhere.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both mention Transformers 5 regressions, but one is about `base_model_prefix` loading and the other about custom initialization; different failure modes and not one concrete bug." + "reason": "Both are Transformers 5.0 loading regressions, but one is about `base_model_prefix` loading and the other about custom model initialization; different failure modes and likely different fixes." }, { "left": "issue:43296", "right": "issue:46821", "accept": false, - "reason": "Completely different models and errors: PaddleOCR-VL load failure vs TimesFM `window_size` AttributeError." + "reason": "Unrelated models and symptoms: PaddleOCR-VL/vLLM loading versus TimesFM `window_size` attribute handling." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Different architectures and bugs: Video-LLaVA tower/weight sharing vs Sam3Model text encoder weights not loading." + "reason": "Different models and root causes: Video-LLaVA temporal attention/weight sharing versus SAM3 text encoder weights not loading." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Both are training/pathology reports, but one is flash-attention-3 degeneracy and the other is a single-label loss edge case across classifiers." + "reason": "Both mention degenerate training/loss behavior, but the concrete triggers differ (`flash_attention_3` regression vs `num_labels=1` classification loss logic)." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Different code paths: non-persistent buffer serialization junk vs `ModelOutput` key assignment behavior." + "reason": "One is a non-persistent buffer serialization bug; the other is `ModelOutput` key assignment behavior. Not the same code-path." }, { - "left": "issue:43818", - "right": "issue:45357", + "left": "issue:42175", + "right": "issue:43296", "accept": false, - "reason": "Both are vision-model issues, but one is Video-LLaVA tower/weights and the other is Qwen3.5 visual encoder key serialization." + "reason": "Packaging/backend dependency issue versus a model loading failure; no shared bug." }, { - "left": "issue:42175", - "right": "issue:43296", + "left": "issue:43818", + "right": "issue:45357", "accept": false, - "reason": "Unrelated packaging/backend issue vs a model-loading failure." + "reason": "Both involve multimodal model weights, but one is Video-LLaVA architecture/weights and the other is Qwen3.5 save_pretrained key serialization." }, { "left": "issue:43493", "right": "issue:43994", "accept": true, - "reason": "Same underlying SigLIP2 model bug: both report incorrect/nonsensical outputs from the HF SigLIP2 implementation, one versus JAX and one in AutoModel/pipeline." + "reason": "Both are about SigLIP2 behaving incorrectly in HF vs expected behavior, and the second looks like a downstream symptom of the same implementation discrepancy." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "Export failure with `torch_compilable_check` is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Different export/inference problems: torch.export with `torch_compilable_check` versus bfloat16 dtype mismatches in specific models." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Different Qwen2.5-VL problems: finetuning with multi-image iterable datasets vs video input position-id regression." + "reason": "Both involve Qwen multimodal inputs, but one is iterable-dataset multi-image finetuning and the other is video position-id handling; not the same bug." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Different bug classes: Qwen2.5-VL video/position-id scaling vs Kosmos2.5 OCR long-input indexing." + "reason": "Different models and failures: Qwen2.5-VL RoPE/position-id scaling versus Kosmos2.5 OCR index errors." }, { "left": "issue:45325", "right": "issue:45381", - "accept": true, - "reason": "Same Qwen2.5-VL visual position-id regression; one identifies the `get_rope_index` scaling bug, the other reports the downstream wrong `vision_position_ids` symptom." + "accept": false, + "reason": "Closely related Qwen2.5-VL position-id bugs, but one is a still-image temporal-position scaling issue and the other is video input vision_position_ids; too different to merge confidently." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "Both touch tied weights/lm_head, but one is a model-specific untied-head bug and the other is device-dependent serialization inconsistency." + "reason": "Both mention tied `lm_head` weights, but one is a model-specific weight-tying failure and the other is device-dependent serialization inconsistency." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Different models and failures: Qwen3Moe FP8 config vs Qwen2.5 GGUF loading under Transformers v5." + "reason": "Different models and failure domains: FP8 quantization on Qwen3Moe versus GGUF loading on Qwen2.5." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to tokenization changes across Transformers versions." + "reason": "Quantization/offload behavior versus tokenization changes; no shared underlying bug." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint key mismatch during resume is unrelated to Qwen2.5-VL position-id generation." + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL position-id computation." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Different loading problems: `init_empty_weights` argument incompatibility vs strict config rejection for `granite_speech`." + "reason": "`init_empty_weights` argument handling and strict config loading for granite_speech are different regressions." }, { "left": "issue:42796", "right": "issue:45381", "accept": false, - "reason": "Tokenizer BPE-marker handling is unrelated to Qwen2.5-VL video position ids." + "reason": "Tokenizer BPE-marker handling is unrelated to Qwen2.5-VL video position-id computation." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "CPU-offload device mismatch is unrelated to `torch.export` failures from `torch_compilable_check`." + "reason": "CPU offload device mismatch in Bark-small versus export failure with `torch_compilable_check`; different subsystems." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "Perplexity on GPT-OSS/WikiText and Whisper empty-transcription generation are unrelated symptoms and code paths." + "reason": "Perplexity regression on GPT-OSS is unrelated to Whisper generation on empty transcription." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sample-rate defaulting is unrelated to PEFT adapter checkpoint loading." + "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are unrelated." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Both are forward/regression bugs, but one is a Qwen video-input regression and the other is a Gemma4 device-map PLE mismatch." + "reason": "Qwen video-input regression and Gemma4 device-map mismatch are separate issues." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Progressive generation with embeddings/past KV is unrelated to a Qwen3 auto-device-map CUDA assert." + "reason": "Progressive generation with `inputs_embeds`/`past_key_values` is unrelated to Qwen3 auto device-map CUDA asserts." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Bark voice preset and progressive generation are entirely different bugs." + "reason": "Bark voice preset handling and progressive generation with embeddings are unrelated bugs." }, { "left": "issue:43828", "right": "issue:46821", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to a TimesFM `window_size` AttributeError." + "reason": "Phi-tiny-MoE autocast dtype mismatch is unrelated to TimesFM `window_size` attribute errors." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 doc-example errors and Qwen2.5-VL multi-image finetuning failures are different models and issues." + "reason": "SigLIP2 docs/example issues and Qwen2.5-VL multi-image finetuning failures are different problems." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes are unrelated to export failures caused by `torch_compilable_check`." + "reason": "Tokenizer crashes on NER/batched padding are unrelated to export behavior with `torch_compilable_check`." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "AyaVision lm_head tying and Bamba-9B-v2 torch.compile SDPA issues are different model-specific bugs." + "reason": "Weight-tying/serialization bug versus Bamba torch.compile SDPA regression; not the same issue." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "CPU offload device mismatch and DAC latent/forward mismatch are unrelated." + "reason": "Both are model inference mismatches, but one is CPU offload device placement and the other is DAC latent/STE forward mismatch." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "Different model-loading failures for different models; no shared bug mechanism." + "reason": "Separate loading failures for different models." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Different device-placement failures in different models; not the same concrete bug." + "reason": "Qwen3 auto device-map CUDA assert and Gemma4 PLE device mismatch are different model/runtime problems." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The two accepted pairs are plausibly duplicate issues from the titles alone (SigLIP2 implementation discrepancy; Qwen2.5-VL video/position-id regression), and the rejected pairs are generally justified by clearly different models, failure modes, or code paths." + "feedback": "The summary is grounded in the packet: issue:43493 is the clearest SigLIP2 root-cause issue, and the 43493/43994 pair is the only one that reasonably looks like the same bug rather than just a related symptom. The other soft-edge rejections are appropriately conservative and consistent with the titles shown." }, "evaluator_used": true, "retried": false, @@ -4539,8 +4521,8 @@ "issue:40990|issue:42947", "issue:43377|issue:44898", "issue:43454|issue:43720", - "issue:43749|issue:44534", "issue:42617|issue:42915", + "issue:43749|issue:44534", "issue:43317|issue:43873", "issue:43572|issue:45759", "issue:43322|issue:46050", @@ -4548,8 +4530,8 @@ "issue:36331|issue:42617", "issue:44387|issue:45704", "issue:41720|issue:44977", - "issue:44466|issue:45357", "issue:36331|issue:42915", + "issue:44466|issue:45357", "issue:43526|issue:43883", "issue:40990|issue:45701", "issue:43450|issue:45701", @@ -4561,220 +4543,220 @@ "issue:44186|issue:45072", "issue:43873|issue:44756" ], - "prepared_review_unit_hash": "6fcc7a3cc7b7723cf73ba3ac60c075f5cf2f1d85a46621e8e2dde403fc9a8fc9", + "prepared_review_unit_hash": "09ca674e0ec8088d4852ed400f1abd74e8d848c106e4d0ba6d72981fbc3ab240", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36989, - "estimated_input_tokens": 9248, - "estimated_eval_tokens": 18752 + "serialized_chars": 36986, + "estimated_input_tokens": 9247, + "estimated_eval_tokens": 18750 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous bug reports; I did not find any soft pair that looks like the same underlying issue, so all candidate duplicates are rejected. Among the set, the strongest standalone issue is the broad tokenizer regression in issue 45920.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue is justified here because the items span unrelated bugs rather than one duplicate family.", + "summary": "This cluster is heterogeneous: most items are distinct Transformers bugs spanning tokenization, loading/offload, quantization, vision, and training-trainer regressions. None of the provided soft pairs look like true duplicates of the same underlying bug/change.", + "confidence": 0.92, + "canonical_issue_reason": "If a single issue must represent the cluster, issue 43873 is the broadest umbrella, but the overall set is too diverse for a strong canonical issue choice.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 is the clearest high-signal report: it describes a broad v5 tokenizer regression affecting multiple models with a concrete symptom and likely reproducible impact.", + "best_issue_reason": "Issue 43873 is the best representative only because it is a broad loading/offload/quantization report, but it is not a good global duplicate anchor for this mixed cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "Different bugs: MIMI batching/padding-mask behavior vs OmDet input-size mismatch." + "reason": "Both are batched-input bugs, but one is about MIMI encoder padding masks and the other about OmDet-Turbo image size mismatches; different code paths and fixes." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "Different code paths: accelerate+deepspeed weight conversion vs missing gradient-checkpointing flag." + "reason": "Both mention Qwen/MoE-style model handling, but one is about accelerate+deepspeed conversion of MoE weights and the other is a missing gradient-checkpointing flag on a trust_remote_code model." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "Both are tokenizer regressions, but one is special-token insertion for a specific model while the other is wrong token IDs across several models; not the same concrete bug." + "reason": "Both are tokenizer regressions, but they affect different models and failure modes: missing BOS/EOS insertion vs wrong token IDs across several models." }, { "left": "issue:46305", "right": "issue:46315", "accept": false, - "reason": "Unrelated functionality: sliding-window masking off-by-one vs tokenizer lookup performance regression." + "reason": "Both are performance/behavior regressions in windowing/token processing, but one is an off-by-one streaming mask bug and the other is slow convert_tokens_to_ids for slow tokenizers." }, { "left": "issue:42796", "right": "issue:46209", "accept": false, - "reason": "Tokenizer/BPE handling issue vs FP8 inference failure on B200; unrelated." + "reason": "One is a tokenizer BPE-marker handling bug for a specific model; the other is an FP8 inference failure on B200 hardware. No shared underlying defect." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "Different failures: tied-weight config bug causing bad generation vs bfloat16 dtype mismatch during inference." + "reason": "One concerns a processor/model input-shape mismatch; the other is dtype mismatches in bfloat16 inference. Too different to merge." }, { "left": "issue:45823", "right": "issue:46821", "accept": false, - "reason": "Different model/runtime issues: device-map forward mismatch vs TimesFM window_size AttributeError." + "reason": "Both involve model/config loading, but one is a Gemma4 device-mapping forward mismatch and the other is a TimesFM window_size attribute error. Different concrete bugs." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Perplexity anomaly on one model is not the same as LoRA gradient-checkpointing ineffectiveness." + "reason": "One reports high perplexity on a model/dataset combination; the other is gradient checkpointing ineffective with LoRA. They do not describe the same bug." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Different model classes and symptoms: MIMI padding-mask discrepancy vs Perceiver resolution handling." + "reason": "Both are batch-related behavior issues, but one is missing padding mask support in MIMI and the other is a Perceiver image-resolution failure. Not the same fix." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "Different problems: weight tying/generation corruption vs packed-weight unpacking during accelerate load." + "reason": "One is a processor shape bug; the other is a BitNet packed-weight unpacking problem during accelerate loading. Different mechanisms and code paths." }, { - "left": "issue:43749", - "right": "issue:44534", + "left": "issue:42617", + "right": "issue:42915", "accept": false, - "reason": "Both touch loading/memory, but one is FSDP CPU RAM-efficient loading and the other is non-persistent buffer initialization; not the same bug." + "reason": "Both are runtime failures, but one is about running 3d_parallel.py and the other is a Qwen3Moe FineGrainedFP8Config issue. No clear shared underlying bug." }, { - "left": "issue:42617", - "right": "issue:42915", + "left": "issue:43749", + "right": "issue:44534", "accept": false, - "reason": "Unrelated failures: 3d_parallel.py runtime issue vs Qwen3MoE FineGrainedFP8Config failure." + "reason": "Both touch loading/runtime behavior, but one is FSDP CPU RAM efficient loading and the other is non-persistent buffers filled with junk in v5. Distinct issues." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both involve offload/quantization broadly, but the reported failure modes and likely code paths differ too much to treat as duplicates." + "reason": "Both mention offload/quantization, but one is device_map=auto failing to load a dequantized model with cpu offload, while the other is a broader quantization offloading complaint. Too broad to treat as one bug." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "Unrelated config regressions: missing pad_token_idx vs improper text_config unwrapping." + "reason": "Both are config/model setup regressions, but one is a missing pad_token_idx after a 5.0 update and the other is AutoModelForCausalLM.from_config not unwrapping text_config for Qwen multimodal configs." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Different lifecycle stages: Llava loading segfault vs sinq quantized model breakage after save/load." + "reason": "Segmentation fault loading Llava Next models is unrelated to sinq quantization breakage after save/load." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Different models and symptoms: CUDA assert under auto device mapping vs segfault during Llava Next loading." + "reason": "Both are model loading failures, but one is a cudaErrorAssert on Qwen3 with auto device mapping and the other is a segmentation fault with Llava Next. Different failure signatures and likely causes." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Trainer API signature change vs inability to run a 3d_parallel script; unrelated." + "reason": "One is a Trainer API signature change for compute_loss; the other is a 3d_parallel.py execution problem. No shared bug." }, { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Both are memory-related, but one is int4 CUDA reserved memory regression and the other is an apex RMSNorm leak; different bugs." + "reason": "Both are memory-related, but one is CUDA reserved-memory growth under int4 quantization and the other is a T5/apex fused RMSNorm leak. Not the same underlying issue." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Different issues around Qwen-family models: device-mapping CUDA assert vs flash-attention generation failure." + "reason": "Both involve Qwen models, but one is a loading/assert failure and the other is flash-attention generation quality/correctness. Different code paths." }, { - "left": "issue:44466", - "right": "issue:45357", + "left": "issue:36331", + "right": "issue:42915", "accept": false, - "reason": "Both concern serialization, but one is tied lm_head weight inconsistency and the other is incorrect visual encoder keys." + "reason": "A Trainer compute_loss signature mismatch is unrelated to Qwen3Moe FP8 configuration failures." }, { - "left": "issue:36331", - "right": "issue:42915", + "left": "issue:44466", + "right": "issue:45357", "accept": false, - "reason": "Trainer loss-signature change vs MoE FP8 loading failure; unrelated." + "reason": "Both concern serialization of model components in v5, but one is tied lm_head.weight serialization and the other is incorrect visual encoder keys for Qwen3.5. Different artifacts and fixes." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "Different components: BeitImageProcessorFast label reduction vs Molmo tied-weights attribute error." + "reason": "One is a BeitImageProcessorFast label-reduction bug; the other is a missing all_tied_weights_keys attribute on MolmoForCausalLM. Unrelated." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "General tokenization/perplexity concerns, but the symptoms and APIs differ; not the same bug." + "reason": "One is a perplexity regression on a specific model/dataset; the other is tokenizer output changing across transformers versions. Not the same defect." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video processor batched-shape bug is unrelated to tokenizer-version behavior." + "reason": "Video processor batch-shape issues are distinct from tokenizer-version-dependent tokenization changes." }, { "left": "issue:43881", "right": "issue:44991", "accept": false, - "reason": "Different failures: model loading for glm-4v-9b vs tokenizer loading for est-roberta." + "reason": "One is a glm-4v-9b loading failure; the other is tokenizer loading failure for a specific RoBERTa model. Different components and symptoms." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "Same model family, but one is a dummy Conv2d in the pixel decoder and the other is missing text encoder weights." + "reason": "A dummy Conv2d in Sam3PixelDecoder is unrelated to Sam3Model text encoder weights not loading." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading failure vs tokenizer loading regression; unrelated." + "reason": "Strict config rejection for granite_speech is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." }, { "left": "issue:26413", "right": "issue:28530", "accept": false, - "reason": "Checkpoint resume/epoch iterator failure vs early stopping metric lookup; different training behaviors." + "reason": "Resume-from-checkpoint epoch iterator failure and early-stopping metric lookup are separate Trainer bugs." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Different failures: CUDA assert during model loading vs a script execution problem." + "reason": "A Qwen3 auto-device-map cuda assert and a 3d_parallel.py run failure are different problems." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash on NER/batched truncation vs dtype mismatch in inference; unrelated." + "reason": "LayoutLMv2Tokenizer NER/padding crashes are unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel inference." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Both mention memory/offload, but quantization offloading and mmap-on-Strix-Halo OOM are not the same underlying bug." + "reason": "Both are memory/offload themed, but one is quantization offloading behavior and the other is disabling mmap on Strix Halo to avoid OOM. Not the same bug." } ] }, @@ -4853,8 +4835,8 @@ "issue:43577|issue:43708", "issue:43295|issue:45701", "issue:44186|issue:44610", - "issue:43550|issue:43606", "issue:45685|issue:46821", + "issue:43550|issue:43606", "issue:43646|issue:46665", "issue:43404|issue:45357", "issue:44610|issue:44898", @@ -4878,214 +4860,208 @@ "issue:43425|issue:46821", "issue:43526|issue:43881" ], - "prepared_review_unit_hash": "94094240657ca8cae67e7f2c19db534c2c5e1aad717fb50c9fccc873b852cbe9", + "prepared_review_unit_hash": "19ff046977a0f742696b51ae9395930d48e8912bd5c43e166bcf7d459f14ea58", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36482, - "estimated_input_tokens": 9121, - "estimated_eval_tokens": 18498 + "serialized_chars": 36487, + "estimated_input_tokens": 9122, + "estimated_eval_tokens": 18500 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: the issues share only broad Transformers/model-loading themes, but not the same concrete bug or change. No soft pair is a safe duplicate merge candidate.", - "confidence": 0.95, - "canonical_issue_reason": "issue:45701 is the best anchor because it has the broadest scope and the most inbound references, so it most plausibly represents the cluster\u2019s general tokenization/regression theme.", + "summary": "This cluster is heterogeneous and mostly contains loosely related Transformers bug reports rather than true duplicates. The strongest sub-theme is generation/cache/compile regressions, with issue 46858 as the best anchor, but there is no single canonical duplicate across the whole set.", + "confidence": 0.29, + "canonical_issue_reason": "Issue 46858 is the closest hub because it is a concrete generation/cache regression and is semantically closest to multiple other cache/compile failures in the cluster, even though the overall set is not a single duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the most central issue to keep as the representative triage target; it is broad enough to absorb related discussion without overclaiming duplicate status.", + "best_issue_reason": "Issue 46858 is the best overall representative: it is specific, technically grounded, and most central to the recurring runtime/generation-cache failures shown by the soft links.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46612", "right": "issue:46858", "accept": false, - "reason": "Both mention generation/cache behavior, but one is about beam-search cache reorder omissions across specific model families while the other is about static-cache seq-length arithmetic on non-CUDA backends; different failures and code paths." + "reason": "Both involve generation internals, but one is beam-search cache reorder logic for legacy models and the other is a static-cache/_prefill device-tensor arithmetic bug in compile paths; not the same fix." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "These are unrelated: SDPA/torch.compile failure in Bamba vs packed-weight unpacking during accelerate loading for BitNet." + "reason": "Different failure modes: torch.compile with SDPA in Bamba versus packed-weight unpacking during accelerate loading for BitNet." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "One is a dtype propagation bug in Blip2 loading; the other is a Trainer resume/max_steps calculation issue." + "reason": "Blip2 dtype propagation on load and Trainer checkpoint-resume step \u0627\u0644\u062d\u0633\u0627\u0628 are unrelated bugs." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Both touch tokenization-related regressions, but one is a specific processor/tokenizer API break in v4.57.5 and the other is a broad version-to-version tokenization change report; not the same concrete bug." + "reason": "Both mention tokenization-related regressions, but one is a processor/tokenizer API break and the other is version-dependent tokenization differences; not the same underlying bug." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Different components and symptoms: LayoutLMv2Tokenizer crashes on NER/padding, while OmDet-Turbo has an image-size mismatch between processor and model." + "reason": "Tokenizer padding/truncation crash in LayoutLMv2 is unrelated to OmDet-Turbo producing the wrong image resolution." }, { - "left": "issue:43550", - "right": "issue:43606", + "left": "issue:45685", + "right": "issue:46821", "accept": false, - "reason": "Both involve device/runtime issues, but the model, mechanism, and failure mode differ completely." + "reason": "An MPS histogram kernel error in MoE code is not the same as a TimesFM window_size AttributeError." }, { - "left": "issue:45685", - "right": "issue:46821", + "left": "issue:43550", + "right": "issue:43606", "accept": false, - "reason": "MPS histogram-on-Int in MoE code is unrelated to TimesFM's window_size AttributeError." + "reason": "Both are runtime issues, but one is a Bamba SDPA compile failure and the other is a CPU offload device-mismatch bug in bark-small." }, { "left": "issue:43646", "right": "issue:46665", "accept": false, - "reason": "One is a Transformers 5.0 custom-model initialization breakage; the other is an inconsistent from_pretrained-in-class behavior report. Similar area, but not the same bug." + "reason": "Custom model initialization breakage and from_pretrained behavior inside Python classes are related only at a very high level, not the same concrete bug." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Weight tying in Mistral3 conditional generation is unrelated to Qwen3.5 save_pretrained visual key serialization." + "reason": "Mistral3 lm_head tying and Qwen3.5 save_pretrained key serialization are different model-state problems." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Both concern image preprocessing, but they are separate model-specific size/positional-encoding problems." + "reason": "Both are vision-model preprocessing issues, but one is a fixed-size processor/model mismatch and the other is positional-encoding interpolation failure." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import error and inability to run 3d_parallel.py are unrelated." + "reason": "An import error for GenerationMixin is unrelated to not being able to run 3d_parallel.py." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "FineGrainedFP8Config failure in Qwen3Moe is unrelated to device_map auto offload loading for a dequantized model." + "reason": "FineGrainedFP8Config failure in Qwen3Moe and dequantized model offload loading failure are different quantization/offload bugs." }, { "left": "issue:44734", "right": "issue:46858", "accept": false, - "reason": "Both involve cache handling, but one is a serve-time KV indexing crash and the other is a seq-length tensor/device arithmetic issue in static cache prefill." + "reason": "Both touch cache continuation, but one is wrong tensor indexing in serve responses and the other is static-cache seq-length device arithmetic; not mergeable as one fix." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Different serve/multimodal processor failures with different root causes and error signatures." + "reason": "Both are serve/multimodal processor regressions, but the concrete errors differ: missing .to() on a string versus Gemma4Processor missing _tokenizer." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Gemma3n variable-name collision in forward is unrelated to Gemma4 PLE device mismatch under device_map=auto." + "reason": "Gemma3n variable-name collision and Gemma4 PLE device mismatch are distinct forward-pass bugs." }, { - "left": "issue:44186", - "right": "issue:44361", - "accept": false, - "reason": "Both are tokenizer/task crashes, but they concern different tokenizers and different code paths." - }, - { - "left": "issue:44464", - "right": "issue:46858", + "left": "issue:46821", + "right": "issue:46869", "accept": false, - "reason": "Chunked-generation inconsistency under compiled forward is not the same bug as static-cache seq-length device tensor arithmetic." + "reason": "TimesFM window_size handling and YOLO multi-GPU device-map failure are unrelated." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collisions for trust_remote_code and read-only file permission propagation are unrelated save/load issues." + "reason": "Cache-path collisions for trust_remote_code and read-only permission propagation in save_pretrained are different filesystem/persistence issues." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Auto device mapping CUDA assert in Qwen3 is unrelated to GPT-OSS MoE tensor-parallel not working." + "reason": "Qwen3 auto-device assert and GPT-OSS MoE tensor-parallel failure are separate model/runtime problems." }, { "left": "issue:28530", "right": "issue:33666", "accept": false, - "reason": "Early stopping metric selection and Qwen2-VL multi-GPU training are unrelated." + "reason": "Early-stopping metric selection and Qwen2-VL multi-GPU training are unrelated training issues." }, { "left": "issue:44464", "right": "issue:46858", "accept": false, - "reason": "Duplicate of the earlier cache/compiled-forward pair: same broad area, but distinct bugs." + "reason": "Both involve compiled generation, but chunked-generation output inconsistency is a different bug than cache sequence-length slice arithmetic on non-CUDA backends." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked-LM prediction regression and Qwen3.5 recurrent-state reset are separate model-specific failures." + "reason": "CamemBERT masked-LM prediction drift and Qwen3.5 recurrent-state reset with cache are different model-specific behaviors." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to StableLmConfig missing pad_token_idx." + "reason": "A sliding-window issue in Qwen3-MoE is unrelated to StableLmConfig missing pad_token_idx after the 5.0 update." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer compute_loss signature change and Qwen3 auto-device-map CUDA assert are unrelated." + "reason": "Trainer compute_loss signature mismatch and Qwen3 device-map CUDA assert are unrelated." }, { "left": "issue:42915", "right": "issue:46291", "accept": false, - "reason": "FineGrainedFP8Config model loading and kernels import/version-revision requirements are unrelated." + "reason": "Qwen3Moe FP8 config failure and kernels package version/revision import requirement are different dependency/runtime issues." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Same as the earlier tokenizer pair: similar symptom class, but different tokenizers and root causes." + "reason": "Both are tokenizer failures, but they affect different tokenizers, tasks, and code paths." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Config API mismatch for Gemma3TextConfig and rotary_pct reset on GPTNeoX reload are separate serialization/API bugs." + "reason": "Gemma3TextConfig API discrepancy and GPTNeoXConfig rotary_pct reload behavior are distinct configuration bugs." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Llava Next segmentation fault are unrelated." + "reason": "Qwen3Moe FP8 loading failure and Llava Next segmentation fault are not the same underlying problem." }, { "left": "issue:43425", "right": "issue:46821", "accept": false, - "reason": "Torch 2.10 compatibility and TimesFM window_size AttributeError are unrelated." + "reason": "Torch 2.10 incompatibility and TimesFM window_size AttributeError are unrelated." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug and glm-4v-9b loading failure are unrelated." + "reason": "BeitImageProcessorFast reduce_labels bug and glm-4v-9b loading failure are different issues." } ] }, @@ -5168,9 +5144,9 @@ "issue:43550|issue:43653", "issue:43606|issue:43720", "issue:44534|issue:46665", - "issue:44492|issue:44509", "issue:41720|issue:43746", "issue:43531|issue:45759", + "issue:44492|issue:44509", "issue:45406|issue:45464", "issue:42490|issue:43329", "issue:43388|issue:45059", @@ -5189,214 +5165,214 @@ "issue:44898|issue:45072", "issue:44625|issue:45479" ], - "prepared_review_unit_hash": "d993255db6252dbd2261ccd7fbcb7f015ae4922dc2b7fbb0f1d221bc4a9d700b", + "prepared_review_unit_hash": "f5ad6ed01f44ee91d298cea32671872e27242f869a5a1244c056914e1bfa1306", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36304, - "estimated_input_tokens": 9076, - "estimated_eval_tokens": 18408 + "serialized_chars": 36305, + "estimated_input_tokens": 9077, + "estimated_eval_tokens": 18410 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are distinct bug reports across unrelated subsystems, so none of the soft-similarity pairs look like true duplicates. If one issue must represent the cluster, the closest recurring theme is the Transformers v5 non-persistent-buffer regression.", - "confidence": 0.43, - "canonical_issue_reason": "issue:44534 is the clearest concrete regression in the set and best matches the recurring Transformers v5 buffer-loading bug theme; it is more representative than the other broadly similar version-regression reports.", + "summary": "This cluster is a false-positive mix of unrelated issues across tokenization, loading, vision/multimodal processing, generation, offloading, and docs. None of the soft edges look like true duplicates, and there are no PRs to compare.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the cluster spans multiple unrelated root causes, so there is no single representative bug to merge everything into.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44534 is the strongest representative issue because it describes a specific, actionable v5 regression with likely broader impact; however, the cluster overall is not a single duplicate family.", + "best_issue_reason": "No issue is a good global canonical; if forced, issue:43606 is the broadest theme (offloading/quantization), but it still does not represent the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:28530", "right": "issue:30064", "accept": false, - "reason": "Early-stopping metric resolution is unrelated to void segmentation-map processing." + "reason": "Early-stopping metric selection is unrelated to image processor handling of void segmentation maps." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are different code paths." + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are different code paths and failures." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "Both are tokenizer bugs, but one is slow/fast inconsistency and the other is a T5 special-tokens load error." + "reason": "Slow/fast tokenizer inconsistency is not the same as a T5Tokenizer load-time add_special_tokens conflict." }, { "left": "issue:44387", "right": "issue:45701", "accept": false, - "reason": "Int4 quantization memory growth is not the same bug as version-dependent tokenization changes." + "reason": "Int4 quantization OOM is a memory/offload problem; tokenization changes are a separate behavior change." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Both involve Qwen multimodal models, but one is a video input ValueError and the other is a weight_only load failure." + "reason": "Both involve Qwen3-family multimodal models, but one is a video-input ValueError and the other is a weight_only loading error." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Quantization/offloading behavior and a T5 apex RMSNorm memory leak are unrelated failures." + "reason": "Quantization/offloading behavior is unrelated to the T5 apex FusedRMSNorm memory leak." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Segmentation-map preprocessing does not match a BarkProcessor voice_preset issue." + "reason": "Void segmentation-map processing and BarkProcessor voice_preset handling are unrelated processor bugs." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "torch.compile SDPA failure and tokenizer special-token registration are unrelated." + "reason": "torch.compile + SDPA on Bamba is unrelated to BigBirdTokenizer special-token registration." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both concern loading/offload infrastructure, but the failing mechanisms are different: CPU offload mismatch vs packed-weight unpacking." + "reason": "Both mention loading/offload, but one is CPU offload with quantization and the other is BitNet packed-weight unpacking during accelerate loading." }, { "left": "issue:44534", "right": "issue:46665", "accept": false, - "reason": "Non-persistent buffer junk in v5 is a different bug from inconsistent from_pretrained behavior inside Python classes." - }, - { - "left": "issue:44492", - "right": "issue:44509", - "accept": false, - "reason": "A typo in cache strategies is unrelated to stale docs for removed pipeline tasks." + "reason": "Non-persistent buffer initialization junk and from_pretrained inside Python classes are distinct issues." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Auto device-mapping cuda assert and PEFT local-checkpoint loading are different issues." + "reason": "Qwen3 auto device mapping cudaErrorAssert and GraniteSpeech PEFT local-checkpoint loading are different failures." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is not the same as unwrapping text_config for composite multimodal configs." + "reason": "Qwen3-MoE sliding_window behavior and AutoModelForCausalLM.from_config text_config unwrapping are not the same bug." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A cache-strategies typo in docs is unrelated to stale docs mentioning removed pipeline tasks." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming chat/completions failure are unrelated." + "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming chat/completions failures are different components and symptoms." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Trailing-slash local model path warnings do not overlap with the undefined video-token counting bug." + "reason": "Trailing-slash local model import warnings are unrelated to the multimodal video-branch undefined-variable bug." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Last-batch label truncation in gather_for_metrics is unrelated to SAM3 text+box behavior." + "reason": "Metric gathering truncation and SAM3 text/bbox weird behavior are separate issues." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "CPU offload/device mismatch and BigBirdTokenizer special-token registration are different bug classes." + "reason": "CPU offload/quantization device mismatch is unrelated to tokenizer special-token registration." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counting and progressive generation with embeds/past keys are not the same defect." + "reason": "TrainerState token-count tracking and progressive generation are unrelated." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "Bark voice_preset loading and Qwen3 auto device mapping crashes are unrelated." + "reason": "BarkProcessor preset handling and Qwen3 auto device mapping/cudaErrorAssert are unrelated." }, { "left": "issue:43531", "right": "issue:46821", "accept": false, - "reason": "Qwen3-MoE sliding-window failure is not the same as TimesFM window_size AttributeError." + "reason": "Qwen3-MoE sliding_window and TimesFM window_size AttributeError are different models and APIs." }, { "left": "issue:43540", "right": "issue:46858", "accept": false, - "reason": "Video-input processing in Qwen3OmniMoe is unrelated to static-cache seq-length tensor arithmetic." + "reason": "Qwen3OmniMoe video ValueError and static-cache seq_length tensor arithmetic are unrelated." }, { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Both are Qwen3.5-related, but one is config unwrapping and the other is CPU-offload incompatibility in rotary embedding." + "reason": "Qwen3.5/Qwen3.6 config unwrapping and CPU-offload rotary embedding compatibility are different problems." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Incorrect visual encoder keys on save_pretrained and missing text-encoder weights on load are different save/load bugs." + "reason": "Incorrect visual-encoder keys on save_pretrained and missing text-encoder weight loading are different load/save regressions." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "A specific model load failure for ScandiBERT is not the same as a tokenizer-loading regression for a different model." + "reason": "Loading ScandiBERT and tokenizer behavior changes across versions are not the same underlying issue." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "An optional post_init policy change is unrelated to num_labels propagation in Qwen3.5 configs." + "reason": "Optional post_init policy for patchtsmixer and Qwen3.5 num_labels propagation are unrelated." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Sliding-window handling and missing Float8 storage support are unrelated." + "reason": "Qwen3-MoE sliding_window and Float8 storage lookup errors are unrelated." }, { "left": "issue:43644", "right": "issue:46665", "accept": false, - "reason": "Non-persistent buffer junk and inconsistent from_pretrained-in-class behavior are distinct issues." + "reason": "Non-persistent buffer junk in Transformers 5.0 and from_pretrained-in-class behavior are unrelated." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "AddedToken special-argument conflict is unrelated to the ScandiBERT load failure." + "reason": "AddedToken special-argument duplication and ScandiBERT loading failure are different tokenizer/model-loading issues." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Perceiver interpolation failures and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are different bugs." + "reason": "Perceiver interpolation on non-default resolution and bfloat16 dtype mismatches are distinct bugs." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "num_labels propagation in Qwen3.5 configs is unrelated to degenerate single-label classification loss." + "reason": "num_labels propagation in Qwen3.5 and degenerate zero loss in sequence classification are not the same code path." } ] }, @@ -5484,11 +5460,11 @@ "issue:44918|issue:44936", "issue:45759|issue:45923", "issue:43746|issue:43873", - "issue:44514|issue:44545", "issue:43421|issue:43653", "issue:43994|issue:44671", - "issue:39692|issue:43493", + "issue:44514|issue:44545", "issue:46032|issue:46463", + "issue:39692|issue:43493", "issue:43421|issue:44186", "issue:45081|issue:45356", "issue:43782|issue:45381", @@ -5502,226 +5478,226 @@ "issue:45987|issue:46829", "issue:43295|issue:45834" ], - "prepared_review_unit_hash": "a5156ed720eb90f3f108563148bbf93222788f5c2650c797370977e8582c3e3d", + "prepared_review_unit_hash": "1ce66ff6d2642937239b05db676cb75233e915188c7b47cd58ca8bb8d9d9161f", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37444, - "estimated_input_tokens": 9361, - "estimated_eval_tokens": 18978 + "serialized_chars": 37450, + "estimated_input_tokens": 9363, + "estimated_eval_tokens": 18982 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous Transformers bug reports with overlapping tokenizer/processor themes, but only one clear duplicate pair stands out: the two Qwen2_5_VLProcessor.apply_chat_template crashes on batched input with padding=False. The rest are related by subsystem at best, not by the same concrete bug.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 44514 is the earliest report in the only exact duplicate pair (44514/44545) and cleanly describes the concrete crash in Qwen2_5_VLProcessor.apply_chat_template with batched input and padding=False.", + "summary": "This cluster is mostly heterogeneous transformer/model bugs with one clear duplicate pair and one likely offline-tokenizer duplicate. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44545 is the cleanest duplicate target for the only exact issue duplication in the set (same Qwen2_5_VLProcessor.apply_chat_template crash as issue:44514).", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44514 is the best representative because it matches its twin exactly and has the clearest canonical repro of the actual failure mode.", + "best_issue_reason": "issue:44545 is the best overall cluster anchor: it is narrowly scoped, clearly reproducible, and has the strongest duplicate signal; the rest of the items are largely unrelated sub-issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Both are training/runtime problems, but they involve different mechanisms and code paths: eval-mode checkpointing vs ZeRO-3 gradient growth." + "reason": "Different bugs: eval-mode gradient checkpointing misuse vs ZeRO-3/deepspeed gradient inflation." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different regressions in different areas: model loading failure vs PIL image processor torchvision dependency." + "reason": "Unrelated failures: a model-loading regression vs PIL image processors incorrectly requiring torchvision." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "One is a warning about tie_word_embeddings during LoRA fine-tuning; the other is a save_pretrained regression for visual encoder keys." + "reason": "Different code paths and symptoms: a tie_word_embeddings warning during LoRA fine-tuning vs incorrect visual encoder keys in save_pretrained." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Both are multimodal-related, but one is a SigLIP2 docs/model mismatch and quantization issue; the other is an OmDet-Turbo processor image-size mismatch." + "reason": "Different models and defects: SigLIP2 doc example errors vs OmDet-Turbo processor/model input-size mismatch." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Completely different problems: Qwen3 device-mapping crash vs T5 apex RMSNorm memory leak." + "reason": "No shared bug: Qwen3 device-mapping CUDA assert is unrelated to T5/apex RMSNorm memory leak." }, { "left": "issue:43746", "right": "issue:46050", "accept": false, - "reason": "Both involve loading/saving models, but the concrete failures differ: PEFT local checkpoint loading vs quantized sinq save/load breakage." + "reason": "Different loading/save problems: PEFT adapter local-checkpoint loading vs sinq quantized model save/load breakage." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Both are tokenization/special-token issues, but one is runtime post-processor refresh and the other is custom token IDs lost after save/load." + "reason": "Both involve tokenizers, but the failures differ: runtime post-processor updates vs config persistence losing custom token IDs." }, { "left": "issue:46655", "right": "issue:46693", "accept": false, - "reason": "Unrelated bugs: video processor padding calculation vs generation performance regression with FA2 and long context." + "reason": "Different subsystems: video-processor padding math vs FA2 long-context generation performance regression." }, { "left": "issue:43381", "right": "issue:45704", "accept": false, - "reason": "Different subsystems and symptoms: gradient checkpointing/eval mode vs T5 RMSNorm memory leak." + "reason": "Completely different bugs in different components." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "No shared bug: T5 memory leak versus a missing supports_gradient_checkpointing flag on a trust_remote_code model." + "reason": "Different issues: apex RMSNorm leak vs missing supports_gradient_checkpointing flag." }, { "left": "issue:42886", "right": "issue:44843", - "accept": false, - "reason": "Both mention HF_HUB_OFFLINE, but one is a generic cache-loading failure and the other is a specific model_info() call in _patch_mistral_regex; not the same concrete bug." + "accept": true, + "reason": "Same underlying offline tokenizer failure: both describe HF_HUB_OFFLINE cache-loading breakage, and 44843 identifies the unconditional model_info() call that explains 42886." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "Different trainer failures: SFT embedding unpacking vs evaluate() failing after train()." + "reason": "Different trainer failures: SFT unpacking embeddings vs evaluate() failing after train()." }, { "left": "issue:45759", "right": "issue:45923", "accept": false, - "reason": "Different issues: unwrapping text_config for composite multimodal configs vs missing gradient-checkpointing support." + "reason": "Different model/config problems: unwrapping text_config vs missing gradient checkpointing flag." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "Different loading/offloading problems: PEFT checkpoint loading versus quantization/offloading behavior." - }, - { - "left": "issue:44514", - "right": "issue:44545", - "accept": true, - "reason": "Exact duplicate reports of the same crash in Qwen2_5_VLProcessor.apply_chat_template with batched input and padding=False." + "reason": "Both mention loading/quantization, but one is PEFT checkpoint loading and the other is offloading with quantization; not the same concrete bug." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Related tokenizer area, but not the same bug: runtime special-token post-processor updates vs BigBird mask token registration/empty decode." + "reason": "Related tokenizer area, but not the same bug: post-processor update logic vs BigBirdTokenizer special-token registration/empty decode." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Different models and failures: SigLIP2 nonsensical outputs vs CamemBERT masked LM prediction regression." + "reason": "Different models and symptom classes: SigLIP2 inference mismatch vs CamemBERT masked-LM prediction regression." }, { - "left": "issue:39692", - "right": "issue:43493", - "accept": false, - "reason": "Both concern SigLIP2, but one is a docs/example mismatch and quantization failure while the other is an implementation discrepancy against JAX." + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Exact duplicate report: same Qwen2_5_VLProcessor.apply_chat_template batched-input crash with padding=False." }, { "left": "issue:46032", "right": "issue:46463", "accept": false, - "reason": "Unrelated: Mamba2 cache correctness bug vs T5 decoder-input error message wording." + "reason": "Unrelated bugs: Mamba2Mixer cache correctness vs T5 decoder input error-message wording." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Same model family, but different problems: doc/example misuse and quantization failure vs a core HF-vs-JAX implementation discrepancy." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Both are tokenizer bugs, but they are different concrete failures on different tokenizers and inputs." + "reason": "Both are tokenizer issues, but the specific failures differ: special-token post-processor updates vs NER/batched padding-truncation crashes." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both involve fix_mistral_regex, but one is an AttributeError crash and the other is a Kimi-K2.5 regression with codec-handling/warning behavior; not clearly the same bug." + "reason": "Both reference Mistral regex handling, but the reported failures are different enough to treat as separate regressions." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Different Qwen VL issues: weight_only loading error vs wrong vision_position_ids for video input." + "reason": "Different Qwen VL bugs: weight_only loading error vs incorrect vision_position_ids for video input." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Completely unrelated: segmentation-map processing vs GenerationMixin import error." + "reason": "No overlap: image-processing segmentation bug vs GenerationMixin import error." }, { "left": "issue:45507", "right": "issue:46858", "accept": false, - "reason": "Different model/code-path failures: GraniteMoEHybrid invalid method call vs cache.get_seq_length() tensor arithmetic issue." + "reason": "Different code paths: invalid method call in GraniteMoEHybrid vs static-cache seq-length tensor breaking slice arithmetic." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Different domains entirely: tokenizer regression vs hub list_repo_files revision forwarding." + "reason": "Unrelated: tokenizer regression in Kimi-K2.5 vs list_repo_files not forwarding revision." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "Only a broad multimodal/quantization overlap; the specific bugs are different." + "reason": "Different problems: SigLIP2 docs/example failures vs quantization/offloading behavior." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "Different model-loading/math bugs: BitNet packed-weight unpacking vs DAC from_latents missing STE." + "reason": "Different model implementations and failures: BitNet packed-weight loading vs DAC from_latents/STE mismatch." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "Different concrete failures in unrelated models: BigBirdTokenizer special-token registration vs BitNet accelerate loading." + "reason": "No shared underlying bug: BigBirdTokenizer special-token handling vs BitNet accelerate-loading of packed weights." }, { "left": "issue:46829", "right": "issue:46832", "accept": false, - "reason": "Different model heads and failure modes: hidden_states output misuse vs RT-DETR feature-level mismatch." + "reason": "Different vision-model bugs: incorrect hidden_states output vs RT-DETR feature-level indexing error." }, { "left": "issue:45987", "right": "issue:46829", "accept": false, - "reason": "Different code paths: cache length return type vs classification output field mapping." + "reason": "Cache API return-type bug vs VideoPrism output-field misassignment." }, { "left": "issue:43295", "right": "issue:45834", "accept": false, - "reason": "Both are multimodal processor issues, but they affect different models and failure modes (processor.tokenizer regression vs Kosmos2.5 OCR index error)." + "reason": "Both are multimodal/processor issues, but they affect different models and different failure modes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the only clearly supported duplicate is issue:44514/issue:44545, and the rest are described conservatively as related-at-best rather than the same concrete bug. The soft-edge verdicts avoid overclaiming overlap and stay aligned with the issue titles/body themes." + "feedback": "The summary is grounded in the packet: it correctly identifies the exact duplicate pair (44514/44545), the likely offline-tokenizer duplicate (42886/44843), and does not claim any PRs exist. The soft-edge verdicts are consistently conservative, with no clear overstatement of evidence." }, "evaluator_used": true, "retried": false, @@ -5792,8 +5768,8 @@ "issue:46932" ], "soft_pairs": [ - "issue:44062|issue:44488", "issue:43550|issue:43927", + "issue:44062|issue:44488", "issue:43550|issue:44361", "issue:43540|issue:46294", "issue:45381|issue:46693", @@ -5805,10 +5781,10 @@ "issue:43653|issue:43819", "issue:41720|issue:44991", "issue:45987|issue:46932", - "issue:43643|issue:44877", - "issue:43618|issue:43697", "issue:30064|issue:36331", + "issue:43643|issue:44877", "issue:44220|issue:44661", + "issue:43618|issue:43697", "issue:43335|issue:43756", "issue:43550|issue:43819", "issue:44843|issue:45081", @@ -5821,202 +5797,202 @@ "issue:45183|issue:45479", "issue:43644|issue:43749" ], - "prepared_review_unit_hash": "3a3401ffcc1b834622537187408bf2607d67820f5d358346a8cfc294fe3aecf0", + "prepared_review_unit_hash": "c77ddd8870b9457b2fe9f7e5d82042246e1b847445dee1d15456d24263fdd32b", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36262, - "estimated_input_tokens": 9066, - "estimated_eval_tokens": 18388 + "serialized_chars": 36259, + "estimated_input_tokens": 9065, + "estimated_eval_tokens": 18386 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a mixed bag of unrelated bug reports across tokenizers, multimodal models, loading/config, and cache behavior. None of the soft pairs look like the same concrete underlying bug or change, so all soft edges should be rejected.", - "confidence": 0.86, - "canonical_issue_reason": "No true duplicate center is evident; issue #43749 is the strongest stand-alone representative because it is a concrete runtime bug with the most discussion and references.", - "canonical_pr_reason": null, - "best_issue_reason": "Issue #43749 is the best cluster representative: it is specific, reproducible, and appears to have the broadest impact/engagement among the listed issues.", - "best_pr_reason": null, + "summary": "This cluster is not a duplicate set: the issues span unrelated bugs across tokenizers, vision/video models, training, config loading, and cache behavior. No safe canonical issue/PR stands out.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits this cluster; the reports describe distinct failures in different components and code paths.", + "canonical_pr_reason": "No PRs are present in the provided items.", + "best_issue_reason": "No global best issue candidate: the items are heterogeneous and not interchangeable as a representative duplicate target.", + "best_pr_reason": "No PRs are present in the provided items.", "soft_edge_verdicts": [ { - "left": "issue:44062", - "right": "issue:44488", + "left": "issue:43550", + "right": "issue:43927", "accept": false, - "reason": "Both are tokenizer-related, but one is an AddedToken API conflict and the other is a model-loading failure; different bugs." + "reason": "Different failures: torch.compile+SDPA on Bamba vs custom token IDs lost after save/load on DiaConfig." }, { - "left": "issue:43550", - "right": "issue:43927", + "left": "issue:44062", + "right": "issue:44488", "accept": false, - "reason": "Torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to DiaConfig losing custom token IDs after save/load." + "reason": "Both are tokenizer-related, but one is an AddedToken keyword conflict and the other is a model-specific load failure." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "These are different code paths and symptoms: Bamba compile+SDPA versus an MLukeTokenizer AttributeError." + "reason": "Unrelated models and root causes: Bamba compile/SDPA vs MLukeTokenizer AttributeError." }, { "left": "issue:43540", "right": "issue:46294", "accept": false, - "reason": "Both involve video inputs, but the failures are different models and different mechanisms; not the same concrete bug." + "reason": "Both involve video inputs, but they affect different Qwen models and different mismatch logic." }, { "left": "issue:45381", "right": "issue:46693", "accept": false, - "reason": "Qwen2.5-VL video position-id corruption is unrelated to a long-context FA2 performance regression." + "reason": "Different subsystems: Qwen2.5-VL video position IDs vs FA2 long-context performance regression." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE and AutoConfig trust_remote_code field loss are unrelated." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to AutoConfig missing fields with trust_remote_code." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Deepspeed ZeRO3/BertModel initialization error and fbank feature extraction issue do not share the same underlying bug." + "reason": "DeepSpeed ZeRO3/BertModel index error is unrelated to the fbank feature extraction issue." }, { "left": "issue:43611", "right": "issue:46665", "accept": false, - "reason": "Both touch loading behavior, but base_model_prefix breakage is distinct from inconsistent from_pretrained behavior in Python classes." + "reason": "Both mention from_pretrained, but the concrete bugs differ: base_model_prefix loading break vs class-context inconsistency." }, { "left": "issue:44008", "right": "issue:46858", "accept": false, - "reason": "Gemma3n variable-name collision and StaticCache sequence-length tensor typing are separate issues." + "reason": "Gemma3n forward variable collision is unrelated to static cache seq_length tensor behavior." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "LayoutLMv2Tokenizer NER/padding crashes and Perceiver image-resolution failure are unrelated." + "reason": "Different models and failures: LayoutLMv2Tokenizer input handling vs Perceiver image classification resolution handling." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration and DAC.from_latents forward mismatch are different components and bugs." + "reason": "Tokenizer special-token registration bug is unrelated to DAC from_latents vs forward mismatch." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "Qwen3 CUDA assert with device mapping is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Qwen3 auto device mapping CUDA assert is unrelated to tokenizer loading for EMBEDDIA/est-roberta." }, { "left": "issue:45987", "right": "issue:46932", "accept": false, - "reason": "StaticCache.get_seq_length() type/shape bug is unrelated to Gemma4 quantized embedder dtype casting." + "reason": "StaticCache seq_length contract bug is unrelated to Gemma4 quantized vision/audio dtype casting." }, { - "left": "issue:43643", - "right": "issue:44877", + "left": "issue:30064", + "right": "issue:36331", "accept": false, - "reason": "Both concern config/loading, but missing fields from trust_remote_code and strict config rejection are distinct failures." + "reason": "ImageProcessor segmentation-map failure is unrelated to CustomTrainer compute_loss argument mismatch." }, { - "left": "issue:43618", - "right": "issue:43697", + "left": "issue:43643", + "right": "issue:44877", "accept": false, - "reason": "CLIPOutput attentions assignment and RTDetrV2 output drift in v5 are different issues." + "reason": "Both are config-loading issues, but one is missing fields with trust_remote_code and the other is strict config rejection for granite_speech." }, { - "left": "issue:30064", - "right": "issue:36331", + "left": "issue:44220", + "right": "issue:44661", "accept": false, - "reason": "Void segmentation map handling in image processing is unrelated to CustomTrainer.compute_loss signature changes." + "reason": "Audio fbank feature extraction and add-new-model-like TOKENIZER_MAPPING_NAMES handling are separate bugs." }, { - "left": "issue:44220", - "right": "issue:44661", + "left": "issue:43618", + "right": "issue:43697", "accept": false, - "reason": "Fbank feature extraction and add-new-model-like failing in TOKENIZER_MAPPING_NAMES are separate bugs." + "reason": "Different output regressions in different models; no shared concrete code-path problem is evident." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Sparse-layer construction in SwitchTransformers and Smollm3 RoPE layer dropping are different model bugs." + "reason": "SwitchTransformers sparse-layer creation and Smollm3 RoPE-layer dropping are unrelated architecture bugs." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Bamba torch.compile/SDPA failure and DAC.from_latents mismatch do not share a code path." + "reason": "torch.compile+SDPA failure on Bamba is unrelated to DAC from_latents/forward mismatch." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Both mention _patch_mistral_regex, but one is an offline hub lookup issue and the other is an attribute error on tokenizers.Tokenizer; not the same defect." + "reason": "Same helper area (_patch_mistral_regex), but different concrete failures: offline model_info call vs backend_tokenizer attribute error." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Pegasus output differences and CamemBERT masked-LM prediction errors are unrelated model-specific regressions." + "reason": "Both are v5 output regressions, but they affect different models and likely different code paths." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "SigLIP2 documentation errors and Kosmos2.5 long-OCR index errors are different problems." + "reason": "SigLIP2 doc example errors are unrelated to Kosmos2.5 long-OCR index error." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "LayoutLMv2Tokenizer padding/NER crash and AutoTokenizer loading FastSpeech2ConformerTokenizer are unrelated tokenizer issues." + "reason": "LayoutLMv2Tokenizer crash is unrelated to AutoTokenizer failing to load FastSpeech2ConformerTokenizer." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor URL loading and template-node compilation errors are different subsystems and bugs." + "reason": "AutoImageProcessor URL loading and template-node compilation are unrelated problems." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError and Gemma3n audio_mel_mask collision are unrelated multimodal bugs." + "reason": "Different Gemma/Qwen code paths and failure modes; both are multimodal but not the same bug." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "BarkProcessor voice_preset failure and GenerationMixin import error are separate issues with no shared concrete code path." + "reason": "BarkProcessor voice_preset and GenerationMixin import error are unrelated issues." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Input validation error messaging and degenerate zero-loss classification behavior are different problems." + "reason": "Both are multi-model behavior changes, but one is input validation messaging and the other is classification loss degeneracy." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer initialization junk and FSDP CPU RAM efficient loading are both loading-related but not the same bug." + "reason": "Non-persistent buffer junk in Transformers 5.0.0 is unrelated to broken FSDP CPU RAM efficient loading." } ] }, @@ -6097,8 +6073,8 @@ "issue:44479|issue:45834", "issue:43550|issue:45072", "issue:43782|issue:44977", - "issue:44561|issue:45020", "issue:43421|issue:44811", + "issue:44561|issue:45020", "issue:43577|issue:43818", "issue:46294|issue:46463", "issue:44991|issue:46291", @@ -6112,25 +6088,25 @@ "issue:43540|issue:45381", "issue:43720|issue:45072", "issue:43825|issue:43827", - "issue:44671|issue:45701", "issue:43645|issue:46665", - "issue:34634|issue:36331", - "issue:43577|issue:43582" + "issue:44671|issue:45701", + "issue:43577|issue:43582", + "issue:34634|issue:36331" ], - "prepared_review_unit_hash": "b08df1f4b992f15c798666e79fdf47a58802c4d97531a36c429eb2ab6976b723", + "prepared_review_unit_hash": "501d8f0033fcf82ebecae9f59cbd77516c424850ba632712de69269d3feea9f4", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35467, + "serialized_chars": 35468, "estimated_input_tokens": 8867, "estimated_eval_tokens": 17990 }, @@ -6139,184 +6115,175 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated transformers bugs; the only plausible duplicate-like pair is the tokenizer inconsistency/additional-space case. The rest differ by model, subsystem, or failure mode.", - "confidence": 0.74, - "canonical_issue_reason": "issue:29159 is the broadest tokenizer-regression report in the small tokenizer-related slice and can represent the added-token spacing symptom in issue:28218; the rest of the cluster is too heterogeneous for a stronger global canonical.", + "summary": "These items are mostly unrelated issue reports across tokenizer, multimodal, loading, dtype, and trainer subsystems. I found no convincing duplicate pairs, so this cluster should not be merged into a single canonical bug.", + "confidence": 0.97, + "canonical_issue_reason": "No true duplicate stands out; if one issue must anchor the cluster, issue 45701 is the broadest tokenizer-regression report and the closest to the small tokenizer-related overlap, but the set is still too heterogeneous.", "canonical_pr_reason": null, - "best_issue_reason": "issue:29159 is the best representative of the tokenizer-related subset, but it is not a good umbrella for the full mixed cluster.", + "best_issue_reason": "Issue 45701 is the most general and potentially umbrella-like of the listed issues, but only by breadth; it is not a strong duplicate canonical because the rest of the cluster spans many distinct failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Empty-transcription behavior after align_special_tokens is unrelated to long-OCR index errors." + "reason": "Different bugs: empty transcription / special-token alignment vs long OCR input index handling." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Llava Next loading segfault and an MPS histogram TypeError are different code paths and failure classes." + "reason": "Unrelated failures: Llava Next load segfault vs MPS histogram TypeError on Int tensors." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "CPU offload device mismatch and bfloat16 dtype mismatch are distinct loading/runtime bugs." + "reason": "Both involve loading/device-type problems, but the concrete failures and models differ too much to be duplicates." }, { "left": "issue:43388", "right": "issue:46032", "accept": false, - "reason": "Last-batch label truncation in gather_for_metrics is unrelated to Mamba2 use_cache correctness." + "reason": "Metric gathering label truncation and Mamba2 cache corruption are separate code paths." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "An error-message wording issue is not the same bug as incorrect SigLIP2 outputs from AutoModel/pipeline." + "reason": "An error-message wording issue is not the same as a model producing nonsensical outputs." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "Qwen video-input regression and OCR long-input indexing error are unrelated." + "reason": "Different model families and different failure modes: video regression vs OCR index error." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "torch.compile SDPA failure in Bamba-9B-v2 is a different issue than generic dtype mismatches." + "reason": "torch.compile+SDPA failure is distinct from bf16 dtype mismatch." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Qwen3VL weight_only loading error and Qwen3.5 flash-attention generation bug are different concrete problems." + "reason": "Weight-only loading error and flash-attention generation bug are not the same underlying issue." }, { - "left": "issue:44561", - "right": "issue:45020", + "left": "issue:43421", + "right": "issue:44811", "accept": false, - "reason": "Both mention remote-code breakage, but one is about removing is_torch_fx_available and the other is broader regression reporting; not the same fixable bug." + "reason": "Both are tokenizer/processor related, but one is post-processor sync at runtime and the other is batch_decode skip_special_tokens handling." }, { - "left": "issue:43421", - "right": "issue:44811", + "left": "issue:44561", + "right": "issue:45020", "accept": false, - "reason": "Runtime post-processor updates for modified special tokens and batch_decode skip_special_tokens handling are different tokenizer behaviors." + "reason": "Both mention trust_remote_code, but one is a missing helper removal and the other is a broader version regression." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "Blip2 dtype propagation and Video-LLaVA temporal-attention/weight-sharing are unrelated model issues." + "reason": "BLIP2 dtype persistence and Video-LLaVA temporal-attention/weight-sharing problems are unrelated." }, { "left": "issue:46294", "right": "issue:46463", "accept": false, - "reason": "Video-token mismatch in Qwen3.5-9B and T5 decoder-input validation messaging are different bugs." + "reason": "Qwen video chat-template mismatch and T5 decoder validation messaging are separate issues." }, { "left": "issue:44991", "right": "issue:46291", "accept": false, - "reason": "Tokenizer loading failure for a specific model and pipelines import failing on kernels version requirements are unrelated." + "reason": "Tokenizer loading failure and kernels dependency import failure are different regressions." }, { "left": "issue:43582", "right": "issue:46512", "accept": false, - "reason": "Apple Silicon warmup TypeError and stale model.config.dtype after from_config are different initialization issues." + "reason": "Apple Silicon allocator TypeError and stale model.config.dtype after from_config are unrelated." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Mistral-3 dtype mismatch and Gemma 3n variable-name collision are not the same code-path problem." + "reason": "dtype mismatch in preprocessing vs variable-name collision AttributeError are not the same bug." }, { "left": "issue:43931", "right": "issue:46846", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and missing linear_attn TP plan entries are related only by model family, not by bug." + "reason": "Both are Qwen3-VL/Qwen3.5 loading issues, but one is a weight-shape mismatch and the other is a TP plan/OOM problem." }, { "left": "issue:28218", "right": "issue:29159", - "accept": true, - "reason": "Both point to tokenizer inconsistency around added-token handling; 29159 is the broader slow-vs-fast mismatch that can cover the extra-space symptom in 28218." + "accept": false, + "reason": "Both are tokenizer complaints, but one is added-token spacing and the other is slow/fast tokenizer inconsistency; not enough evidence of a single root cause." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Mask-shape indexing error and trainer.evaluate() failing after train() are different failures." + "reason": "Mask-shape IndexError and train-then-evaluate failure do not describe the same code-path problem." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Processor/model image-size mismatch and dtype mismatch are separate issues." + "reason": "Image-size mismatch in a processor is unrelated to bf16 dtype mismatch." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "FSDP CPU RAM efficient loading and init_empty_weights unexpected-argument errors stem from different loading mechanisms." + "reason": "FSDP RAM-efficient loading and init_empty_weights _is_hf_initialized argument errors are distinct loading regressions." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Qwen3Omni video processing ValueError and Qwen2.5-VL vision_position_ids regression are different video-input bugs." + "reason": "Both are video-input bugs in Qwen models, but they target different models and different failure mechanisms." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is unrelated to bfloat16 dtype mismatches." + "reason": "Packed-weight unpacking during accelerate loading is a different issue from dtype mismatches." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "One is an incorrect runtime error message, the other is stale docs referencing pipeline(); not the same bug." - }, - { - "left": "issue:44671", - "right": "issue:45701", - "accept": false, - "reason": "CamemBERT masked-LM regression and a broad tokenization-change report are too different to merge." + "reason": "Incorrect error messaging and outdated docs references are separate changes." }, { "left": "issue:43645", "right": "issue:46665", "accept": false, - "reason": "Custom models in Jupyter notebooks and from_pretrained inside Python classes are different context/initialization issues." + "reason": "Notebook custom-model initialization and from_pretrained inside Python classes are related only at a high level, not the same bug." }, { - "left": "issue:34634", - "right": "issue:36331", + "left": "issue:44671", + "right": "issue:45701", "accept": false, - "reason": "Bark voice_preset failure and CustomTrainer compute_loss signature breakage are unrelated." + "reason": "CamemBERT masked-LM regression is not the same as a generic tokenization change report." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "Blip2 dtype propagation and Apple Silicon allocator warmup TypeError are unrelated." + "reason": "BLIP2 dtype state and Apple Silicon allocator warmup are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is well-grounded in the packet: it treats the cluster as mostly heterogeneous and only groups the clearly related tokenizer pair (28218/29159). The soft-edge verdicts are conservative overall, and the rejection reasons track the issue titles without overstating a shared bug across unrelated model/subsystem failures." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -6397,8 +6364,8 @@ "issue:43550|issue:44912", "issue:43452|issue:44991", "issue:43257|issue:43329", - "issue:43329|issue:45600", "issue:43526|issue:43697", + "issue:43329|issue:45600", "issue:44060|issue:44368", "issue:43906|issue:46209", "issue:36010|issue:36331", @@ -6411,202 +6378,202 @@ "issue:42175|issue:46291", "issue:42490|issue:43257" ], - "prepared_review_unit_hash": "f784058c245047c810c9c86a79835540b0422810d0c81c2830f4add5613acd7a", + "prepared_review_unit_hash": "1333573b8ce8c764c792deb91c7954ab829689b07aa824ee1d20170f25a78369", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36024, - "estimated_input_tokens": 9006, - "estimated_eval_tokens": 18268 + "serialized_chars": 36025, + "estimated_input_tokens": 9007, + "estimated_eval_tokens": 18270 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft pairs mostly connect unrelated Transformers bugs across import/loading, training, vision/audio processors, quantization, and model-specific regressions. None of the pairs look like the same underlying issue closely enough to deduplicate.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue cleanly represents this cluster; the items are largely unrelated. If forced to pick the broadest umbrella, #45020 ('Recent transformers versions break models using `remote_code`') is the most general, but it does not subsume the other reports.", + "summary": "These are mostly independent Transformers issues spanning unrelated subsystems; none of the soft pairs look like true duplicates or mergeable PR-equivalents.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue: the cluster mixes unrelated bugs across import, loading, tokenizer, processor, model, and training paths, so one issue would not represent the group well.", "canonical_pr_reason": null, - "best_issue_reason": "#45020 is the best global issue candidate only because it is broad and central to a class of loading regressions, with some inbound references. It is still not a good canonical for the whole set because the cluster contains many distinct, model-specific or subsystem-specific bugs.", + "best_issue_reason": "No global best issue stands out; the items are heterogeneous, so any single issue would be an arbitrary representative rather than a true canonical.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Different problems: one is TensorFlow backend packaging, the other is a Mistral tokenizer regex/path crash." + "reason": "Different problems: backend packaging/dependency selection versus a Mistral regex patch crash during tokenizer loading." }, { "left": "issue:46175", "right": "issue:46932", "accept": false, - "reason": "Both are Gemma4-related, but one is a huge vision patch-embedder memory issue and the other is dtype casting for quantized embedders; not the same bug." + "reason": "Both are Gemma4-related, but one is a huge one-hot materialization issue and the other is quantized embedder dtype casting; different code paths." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Both touch model internals, but one is unsafe sys.modules access and the other is a CLIP class structure inconsistency." + "reason": "sys.modules access in modeling_utils is unrelated to the CLIP text-model structure regression." }, { "left": "issue:46697", "right": "issue:46728", "accept": false, - "reason": "Both concern image processors and padding/patch counting, but they affect different models and different logic paths." + "reason": "Different model processors and symptoms: Swin2SR over-padding versus Aria patch-count miscalculation." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "BigBirdTokenizer special-token decoding and MXFP4 quantization fallback are unrelated." + "reason": "Tokenizer special-token registration bug versus MXFP4 quantization fallback; different subsystems." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Qwen3VL weight_only loading and Qwen2.5-VL rope indexing are separate model regressions." + "reason": "Qwen3VL weight-only loading error is separate from Qwen2.5-VL rope index scaling." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio processor sampling-rate defaulting and DAC latents/STE mismatch are different code paths and symptoms." + "reason": "Audio chat-template sample-rate defaulting is unrelated to DAC latent decoding mismatch." }, { "left": "issue:45372", "right": "issue:46326", "accept": false, - "reason": "Both involve mistral_common, but one is an import dependency break and the other is a continue_final_message behavior bug." + "reason": "Both involve mistral_common, but one is a missing import breaking Gemma4 loading and the other is a continue_final_message tokenizer bug." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Mistral tokenizer regex handling and Gemma4 processor missing _tokenizer are distinct failures." + "reason": "Different failures: tokenizer regex patch crashes versus Gemma4Processor missing _tokenizer in serve." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Apple Silicon warmup TypeError and CUDA float16 AMP crash are different backend/device-specific bugs." + "reason": "Apple Silicon TypeError in caching warmup is unrelated to CUDA AMP loss crashes." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "MobileLLM key error and remote_code loading breakages are not the same concrete defect." + "reason": "A specific MobileLLM loading key error is not the same as the broader remote_code regression." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Float8 storage lookup failure and PIL backend torchvision dependency regression are unrelated." + "reason": "Float8 storage lookup failure is unrelated to PIL image processors requiring torchvision." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "Bamba-9B-v2 torch.compile/SDPA failure is a model/runtime correctness bug, not a quantization loading fallback issue." + "reason": "torch.compile SDPA failure on Bamba is a different bug from MXFP4 quantized loading fallback." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer/model loading failures, but for different causes and different models." + "reason": "gguf_file handling for from_pretrained is unrelated to the est-roberta tokenizer load regression." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3 MoE loading with accelerate+deepspeed and undefined video-token helper variables are separate issues." + "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed is different from an undefined-variable bug in multimodal token counting." }, { - "left": "issue:43329", - "right": "issue:45600", + "left": "issue:43526", + "right": "issue:43697", "accept": false, - "reason": "A multimodal token counting bug and stale auto_mappings config references are unrelated." + "reason": "Beit reduce_labels returning one label is unrelated to RTDetrV2 output drift." }, { - "left": "issue:43526", - "right": "issue:43697", + "left": "issue:43329", + "right": "issue:45600", "accept": false, - "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output differences are distinct model/processor bugs." + "reason": "Undefined video-branch variables in multimodal token counting do not match removed auto-mapping config references." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "The first is a tied-weights warning bug in Qwen3-Next; the second is a tie_word_embeddings warning on Qwen3.5-27B. Related theme, but not the same defect." + "reason": "Both mention tied-weights warnings, but one is an incorrect internal tie and the other is a LoRA/config warning about tie_word_embeddings." }, { "left": "issue:43906", "right": "issue:46209", "accept": false, - "reason": "An isolated reproduction of an older issue and an FP8 inference-on-B200 bug are not the same code-path problem." + "reason": "An isolated reproduction of another issue is not the same as the FP8 inference failure on B200." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "GenerationMixin import error and CustomTrainer.compute_loss signature breakage are unrelated APIs." + "reason": "GenerationMixin import failure and CustomTrainer compute_loss signature mismatch are distinct API breakages." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Beit label reduction and MPS attention correctness are unrelated." + "reason": "ImageProcessor label reduction bug is unrelated to the MPS attention correctness issue." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "GenerationMixin import failure and Jupyter custom model initialization breakage do not share a concrete root cause." + "reason": "ImportError for GenerationMixin is not the same as the Jupyter custom-model initialization breakage." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Template compilation error and CLIPTokenizer model_max_length issue are unrelated." + "reason": "Template compilation TypeError and CLIPTokenizer model_max_length are unrelated tokenizer/model config issues." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC.from_latents/STE mismatch and DiaConfig custom token ID persistence are different model-specific bugs." + "reason": "DAC forward/from_latents mismatch is unrelated to DiaConfig losing custom token IDs after save/load." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "DeepSpeed ZeRO-3 empty tensor IndexError and trainer.evaluate() after train() are separate training workflow issues." + "reason": "deepspeed zero3 IndexError with a non-pretrained Bert model is different from trainer.evaluate() failing after train()." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and PIL backend torchvision dependency are unrelated." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to PIL backend torchvision requirements." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "TensorFlow missing from torch extra install and transformers.pipelines kernels dependency requirements are unrelated packaging problems." + "reason": "Package backend selection for torch installs is unrelated to kernels v0.15 breaking pipelines import." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Trailing-slash local model import warning and Qwen3 MOE weight conversion with accelerate+deepspeed are not the same issue." + "reason": "Trailing-slash local-path import warning is unrelated to Qwen3 MoE weight conversion under accelerate/deepspeed." } ] }, @@ -6642,7 +6609,9 @@ "issue:43262", "issue:43329", "issue:43450", + "issue:43454", "issue:43540", + "issue:43577", "issue:43606", "issue:43643", "issue:43650", @@ -6666,11 +6635,9 @@ "issue:44821", "issue:44912", "issue:44977", - "issue:44987", "issue:45042", "issue:45356", "issue:45459", - "issue:45538", "issue:45759", "issue:45803", "issue:45859", @@ -6682,16 +6649,16 @@ "issue:43720|issue:44610", "issue:43746|issue:44811", "issue:43854|issue:44977", - "issue:39692|issue:44912", "issue:43643|issue:45759", + "issue:39692|issue:44912", "issue:38175|issue:44610", "issue:43257|issue:43696", "issue:43262|issue:44811", "issue:43257|issue:43931", "issue:44387|issue:44617", "issue:43329|issue:43450", - "issue:45459|issue:45803", "issue:43329|issue:44479", + "issue:45459|issue:45803", "issue:46421|issue:46424", "issue:34567|issue:40990", "issue:42915|issue:43742", @@ -6705,198 +6672,198 @@ "issue:43927|issue:44912", "issue:28282|issue:29870", "issue:44568|issue:45356", - "issue:44987|issue:45538" + "issue:43454|issue:43577" ], - "prepared_review_unit_hash": "5113945dc623abb8e4051b08852756a471bfbf365143730b72be67a2f6cf844e", + "prepared_review_unit_hash": "a220334d26285d2dd06c841dad84970622cdf190cbf1205602e0bba569a4c6c5", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35354, - "estimated_input_tokens": 8839, - "estimated_eval_tokens": 17934 + "serialized_chars": 35448, + "estimated_input_tokens": 8862, + "estimated_eval_tokens": 17980 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is heterogeneous: it spans model loading, tokenizer, multimodal/video, quantization, cache, and offload regressions. None of the soft-paired items look like true duplicates of the same concrete bug/change, so I would not merge any pair.", - "confidence": 0.71, - "canonical_issue_reason": "issue:43749 is the best single anchor because it is a concrete loading regression with strong discussion/inbound references, but the overall cluster is too mixed to treat as one duplicate family.", + "summary": "The items are a heterogeneous mix of unrelated bug reports across loading, tokenization, multimodal processing, quantization, and memory issues. None of the proposed soft pairs look like the same underlying bug/change, so no duplicate merges are justified.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43749 is the strongest single anchor in the set: it has the highest discussion/inbound-reference signal and describes a concrete loading-path regression. That said, the overall set is not a true duplicate cluster, so this is only a pragmatic representative.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43749 is the most representative/high-signal issue in this set, but only as a broad loading-regression anchor; it is not a true duplicate umbrella for the others.", + "best_issue_reason": "issue:43749 is the best issue to surface as the representative artifact because it is well-scoped, high-signal, and more broadly actionable than the other reports in this mixed set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Different failures: packed-weight loading in BitNet vs processor/model input-size mismatch in OmDet-Turbo." + "reason": "Different failures: BitNet packed-weight unpacking during accelerate loading vs OmDet-Turbo processor image-size mismatch." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "Different code paths: PEFT/local checkpoint loading vs Whisper processor batch_decode skip_special_tokens handling." + "reason": "Different code paths: PEFT adapter checkpoint loading vs Whisper processor batch_decode skip_special_tokens handling." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Different model-specific bugs: GLM-4.7 unit-test loading issue vs Qwen3.5 flash-attention generation regression." + "reason": "Different models and symptoms: GLM-4.7 unit-test loading failure vs Qwen3.5 flash-attention generation issue." }, { - "left": "issue:39692", - "right": "issue:44912", + "left": "issue:43643", + "right": "issue:45759", "accept": false, - "reason": "Different problem classes: documentation/example errors vs MXFP4 quantization fallback during model loading." + "reason": "Both involve config/from_pretrained behavior, but one is missing fields with trust_remote_code and the other is AutoModelForCausalLM.from_config not unwrapping text_config for composite Qwen multimodal configs." }, { - "left": "issue:43643", - "right": "issue:45759", + "left": "issue:39692", + "right": "issue:44912", "accept": false, - "reason": "Both involve config loading, but one is missing fields from trust_remote_code and the other is AutoModelForCausalLM.from_config not unwrapping text_config for multimodal Qwen configs." + "reason": "Different topics: SigLIP2 doc example errors vs gpt-oss-20b MXFP4 quantization fallback." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "Different symptoms and components: zero probabilities in SigLIP2 vs image processor producing the wrong resolution for OmDet-Turbo." + "reason": "Different bug classes: unexpected zero probabilities in SigLIP2 vs processor/model input-size mismatch in OmDet-Turbo." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Different root causes: Qwen3 MOE weight conversion under accelerate/deepspeed vs GPT-oss-20b CUDA OOM." + "reason": "Qwen3 MoE weight conversion with accelerate+deepspeed is unrelated to GPT-oss-20b CUDA OOM." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sampling-rate default bug vs Whisper batch_decode ignoring skip_special_tokens; unrelated behavior." + "reason": "Audio chat-template sampling-rate default mismatch is unrelated to Whisper batch_decode special-token skipping." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Different Qwen model loading issues: MOE weight conversion failure vs VL-30B weight-shape mismatch." + "reason": "Both are Qwen-related loading problems, but one is MoE weight conversion under accelerate/deepspeed and the other is a weight-shape mismatch for Qwen3-VL-30B." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "One is int4 quantization memory growth, the other is Sam3Video OOM; same symptom class but not the same bug." + "reason": "Different memory symptoms and causes: int4 reserved-memory regression in Transformers 5.x vs Sam3Video CUDA OOM." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Both are video-related, but one is an undefined-token-count helper bug and the other is batched video processor output shape." + "reason": "Video-related only at a broad level; one is undefined variables in multimodal token counting, the other is incorrect batched video processor output shape." }, { - "left": "issue:45459", - "right": "issue:45803", + "left": "issue:43329", + "right": "issue:44479", "accept": false, - "reason": "Both hide errors during conversion, but one is protobuf decode error handling and the other is a bare except in Fuyu tensor conversion." + "reason": "Both touch video/VLM processing, but the bugs are distinct: undefined helper/merge_size in token counting vs a v5.3.0 video-input regression across several models." }, { - "left": "issue:43329", - "right": "issue:44479", + "left": "issue:45459", + "right": "issue:45803", "accept": false, - "reason": "Shared video-regression theme, but different affected models and code paths; not the same underlying bug." + "reason": "Both concern error handling, but one is protobuf decode error masking in tokenizers and the other is a bare except in Fuyu tensor conversion." }, { "left": "issue:46421", "right": "issue:46424", "accept": false, - "reason": "Both concern cache configuration, but one is lazy init causing recompilation and the other is max_cache_length only being honored for Quantized Cache." + "reason": "Both mention cache behavior, but one is lazy initialization causing recompilation and the other is max_cache_length being honored only for Quantized Cache." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Trainer token counting bug vs GPT-oss-20b high perplexity; unrelated." + "reason": "Training token-count tracking bug vs GPT-oss-20b high-perplexity evaluation issue; unrelated failures." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Different loading failures: FineGrainedFP8Config on Qwen3Moe vs key error loading MobileLLM-125M." + "reason": "Both involve model loading/configuration, but one is Qwen3Moe FineGrainedFP8Config failure and the other is a MobileLLM key error." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Both involve memory/loading behavior, but FSDP CPU RAM efficient loading is not the same issue as disabling mmap on Strix Halo to avoid OOM." + "reason": "Both are memory-related, but one is FSDP_CPU_RAM_EFFICIENT_LOADING breakage and the other is mmap causing OOM on Strix Halo." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "Init-empty-weights argument incompatibility is unrelated to PIL backend processors requiring torchvision." + "reason": "Different regression types: init_empty_weights/_is_hf_initialized TypeError vs PIL backend processors wrongly requiring torchvision." }, { "left": "issue:29159", "right": "issue:34567", "accept": false, - "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState token counting." + "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState token-count updates." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "No meaningful technical overlap: the first is an invalid/non-technical issue, the second is a reproduction of another transformers bug." + "reason": "No concrete shared bug: one title is non-informative, the other is an isolated reproduction of a previously reported issue." }, { "left": "issue:43606", "right": "issue:45859", "accept": false, - "reason": "Both mention CPU offload, but they affect different models and code paths; not safe to collapse." + "reason": "Both mention CPU/offload-like behavior, but one is a device mismatch in bark-small CI and the other is a Qwen3_5Moe rotary embedding CPU-offload incompatibility." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "Quantization/config loading failure vs AutoImageProcessor URL loading issue; unrelated." + "reason": "Qwen3Moe FineGrainedFP8Config failure is unrelated to AutoImageProcessor URL loading." }, { "left": "issue:43540", "right": "issue:46701", "accept": false, - "reason": "Different multimodal/video problems: Qwen3OmniMoe video processing ValueError vs DiffusionGemma image-text-to-text pipeline failure." + "reason": "Different multimodal/video pipeline bugs: Qwen3OmniMoe video input ValueError vs DiffusionGemmaGenerationOutput pipeline failure." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "Custom token IDs lost on save/load is unrelated to MXFP4 quantization fallback during load." + "reason": "DiaConfig custom-token persistence bug is unrelated to gpt-oss-20b MXFP4 quantization fallback." }, { "left": "issue:28282", "right": "issue:29870", "accept": false, - "reason": "Torch import error is unrelated to Bart generate decoder_attention_mask UnboundLocalError." + "reason": "ImportError for missing PyTorch is unrelated to Bart generate() decoder_attention_mask unbound-variable error." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Different tokenizer regressions: missing BOS/EOS insertion vs Kimi-K2.5 codec handling and warning behavior." + "reason": "Different tokenizer regressions: add_special_tokens not adding BOS/EOS vs Kimi-K2.5 codec-handling and warning regression." }, { - "left": "issue:44987", - "right": "issue:45538", + "left": "issue:43454", + "right": "issue:43577", "accept": false, - "reason": "Loading failure for physical-intelligence/fast is unrelated to CLIPTokenizer model_max_length overflow." + "reason": "AyaVision lm_head weight tying bug is unrelated to BLIP2 dtype propagation to model and qformer." } ] }, @@ -6928,13 +6895,13 @@ "issue:33357", "issue:34567", "issue:34689", + "issue:36331", "issue:41720", "issue:42175", "issue:42915", "issue:43404", "issue:43425", "issue:43452", - "issue:43454", "issue:43525", "issue:43540", "issue:43577", @@ -6970,7 +6937,7 @@ "issue:46869" ], "soft_pairs": [ - "issue:43454|issue:43577", + "issue:44987|issue:45538", "issue:28282|issue:34689", "issue:43525|issue:46291", "issue:43697|issue:45357", @@ -6983,225 +6950,235 @@ "issue:42915|issue:45538", "issue:42175|issue:43824", "issue:28282|issue:29127", - "issue:44062|issue:44991", "issue:44479|issue:46693", + "issue:44062|issue:44991", "issue:43742|issue:44291", "issue:42915|issue:43749", "issue:44936|issue:46821", "issue:24643|issue:29159", "issue:43452|issue:43582", - "issue:44849|issue:45701", "issue:43577|issue:46512", + "issue:44849|issue:45701", "issue:44987|issue:45907", "issue:27957|issue:28282", "issue:41720|issue:43540", "issue:43540|issue:44560", "issue:44479|issue:45381", - "issue:44568|issue:45488" + "issue:44568|issue:45488", + "issue:34567|issue:36331" ], - "prepared_review_unit_hash": "9e3dc7a2f95e53ad2b596b362f85e067f2707ef4b0a438cb240232c6f2b1f718", + "prepared_review_unit_hash": "7cf32ece9642dcf00d943c0e0d1e66a753ec1609fd039cc811642ac0cde5da91", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36054, - "estimated_input_tokens": 9014, - "estimated_eval_tokens": 18284 + "soft_pair_count": 29, + "serialized_chars": 36367, + "estimated_input_tokens": 9092, + "estimated_eval_tokens": 18440 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a loose collection of unrelated Transformers regressions and model-specific bugs. None of the suggested soft pairs look like the same underlying issue, so I would not merge any of them.", - "confidence": 0.82, - "canonical_issue_reason": "issue:45488 is the strongest representative: it is a concrete, high-activity v5 tokenizer regression with broad user impact and clear repros.", + "summary": "Mostly unrelated issue set. One likely duplicate pair exists around the Qwen2.5-VL video regression (44479/45381); the rest of the soft matches are distinct bugs across different models/components and should stay separate.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44479 is the best representative of the only accepted duplicate pair: it is the broader, earlier report for the Qwen video-input regression, while issue:45381 is a narrower symptom of the same 5.3.0 bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45488 is the best global triage anchor because it is the most broadly scoped and well-described regression in the set.", + "best_issue_reason": "issue:44479 most clearly captures the shared Qwen2.5-VL video regression and is the most general surviving issue in the only real duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43454", - "right": "issue:43577", + "left": "issue:44987", + "right": "issue:45538", "accept": false, - "reason": "Different models and different failure modes: untied lm_head weights vs dtype propagation staying float32." + "reason": "Different failures: model loading regression vs CLIPTokenizer max-length bug." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Both are loading failures, but one is a missing PyTorch dependency error and the other is a model-specific loading regression." + "reason": "Both are loading errors, but they concern different models and different root causes." }, { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Unrelated bugs: missing pad_token_id on Llama4Config vs kernels package revision/version handling." + "reason": "Unrelated issues: missing pad_token_id on Llama4Config vs kernels revision/version validation." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "Different subsystems and symptoms: RT-DETR output mismatch vs Qwen3.5 save_pretrained key naming." + "reason": "Different problems: RTDetrV2 output changes vs Qwen3.5 save_pretrained key serialization." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "Completely unrelated: MacOS bus error in CLIP loading vs TrainerState token counter not updating." + "reason": "MacOS bus error in CLIP loading is unrelated to TrainerState.num_input_tokens_seen not updating." }, { "left": "issue:44857", "right": "issue:46832", "accept": false, - "reason": "Different detection model crashes in different code paths; not the same underlying bug." + "reason": "Related RT-DETR family, but different concrete bugs and code paths." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both mention tied weights, but one is a missing tie in a specific model class and the other is LoRA merge collapse with extended vocab." + "reason": "One is a weight-tying bug; the other is LoRA merge collapse with extended vocab." }, { - "left": "issue:43425", + "left": "issue:43452", "right": "issue:46291", "accept": false, - "reason": "Torch version incompatibility vs kernels revision/version requirement are unrelated dependency issues." + "reason": "Different subsystems: GGUF file handling vs pipelines import-time kernel revision checks." }, { "left": "issue:45538", "right": "issue:46869", "accept": false, - "reason": "CLIP tokenizer max length bug is unrelated to YOLO multi-GPU device mapping failure." + "reason": "CLIPTokenizer max_length bug is unrelated to YOLO multi-GPU device-map failures." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error and unreleased PEFT version bump are unrelated." + "reason": "DeepSpeed ZeRO3 deque error is unrelated to the PEFT minimum-version bump." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Different model/runtime failure vs tokenizer configuration constant; no shared code path." + "reason": "Qwen3Moe FP8 failure and CLIPTokenizer max_length are unrelated." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Packaging/backend omission vs missing model export/import symbol; not the same bug." + "reason": "Packaging/backend dependency issue vs missing Qwen2_5_VL import are different bugs." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "ImportError about PyTorch availability is unrelated to LayoutLMv3 box-validation messaging." + "reason": "AutoModel PyTorch import error is unrelated to the layoutlmv3 box-validation message quality issue." }, { - "left": "issue:44062", - "right": "issue:44991", + "left": "issue:44479", + "right": "issue:46693", "accept": false, - "reason": "Different tokenizer regressions with different triggers; not enough evidence of one underlying defect." + "reason": "Both are regressions, but one is a Qwen video-input bug and the other is a FlashAttention2 long-context performance regression." }, { - "left": "issue:44479", - "right": "issue:46693", + "left": "issue:44062", + "right": "issue:44991", "accept": false, - "reason": "Video-input regression and FA2 long-context performance regression are distinct issues." + "reason": "AddedToken keyword collision is unrelated to tokenizer loading for EMBEDDIA/est-roberta." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "KeyError on model loading vs unexpected _is_hf_initialized arg in init_empty_weights are different failures." + "reason": "Different loading failures: MobileLLM key error vs init_empty_weights unexpected _is_hf_initialized argument." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Qwen3Moe FP8 failure and FSDP CPU RAM efficient loading breakage are unrelated." + "reason": "Qwen3Moe FP8 config failure is unrelated to FSDP CPU RAM efficient loading." }, { "left": "issue:44936", "right": "issue:46821", "accept": false, - "reason": "Trainer evaluate/train lifecycle bug vs TimesFM window_size AttributeError; different APIs and code paths." + "reason": "Trainer evaluate/train interaction is unrelated to TimesFM window_size attribute errors." }, { "left": "issue:24643", "right": "issue:29159", "accept": false, - "reason": "DeepSpeed weight-dimension training error is unrelated to tokenizer slow/fast inconsistency." + "reason": "DeepSpeed training weight-shape error and slow/fast tokenizer inconsistency are different bugs." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "gguf_file loading breakage and Apple Silicon allocator warmup TypeError are unrelated." + "reason": "GGUF loading and AppleSilicon caching_allocator_warmup TypeError do not share the same failure path." }, { - "left": "issue:44849", - "right": "issue:45701", + "left": "issue:43577", + "right": "issue:46512", "accept": false, - "reason": "Qwen3.5 hidden-states bug and broad tokenization-version behavior change are not the same concrete bug." + "reason": "Both involve dtype, but one is an actual Blip2 model-casting bug and the other is stale config.dtype after from_config." }, { - "left": "issue:43577", - "right": "issue:46512", + "left": "issue:44849", + "right": "issue:45701", "accept": false, - "reason": "Both involve dtype state, but one is incorrect model weights and the other is stale config metadata." + "reason": "Qwen3.5 hidden-states bug is unrelated to a generic tokenization-version change report." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Model loading failure and list_repo_files revision forwarding are unrelated." + "reason": "Model-loading failure is unrelated to list_repo_files needing to forward revision." }, { "left": "issue:27957", "right": "issue:28282", "accept": false, - "reason": "Flash Attention 2 on XLMRoberta and missing PyTorch dependency ImportError are different problems." + "reason": "XLMRoberta Flash Attention 2 behavior is unrelated to AutoModel PyTorch import errors." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "Auto device-map CUDA assert and Qwen3OmniMoe video processing ValueError are unrelated." + "reason": "Different Qwen failures: auto device-map cudaErrorAssert vs video-input ValueError in Qwen3OmniMoe." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Both are video-related, but they affect different models and different failure mechanisms." + "reason": "Both mention video inputs, but they are different models and distinct error paths." }, { "left": "issue:44479", "right": "issue:45381", - "accept": false, - "reason": "Both involve Qwen VL video inputs, but one is a broad regression and the other is a specific position-id bug; not clearly the same underlying change." + "accept": true, + "reason": "Same underlying Qwen2.5-VL video-input regression in transformers 5.3.0; 45381 is a narrower symptom of 44479." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer regressions in v5, but the affected tokenizer classes and root causes differ." + "reason": "Tokenizer special-token addition bug and LlamaTokenizer ByteLevel-to-Metaspace override are different regressions." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token-count update bug is unrelated to CustomTrainer.compute_loss signature change." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it identifies the single accepted duplicate-like pair (44479/45381) using matching Qwen video-input regression language, and the remaining soft rejections are phrased as distinct bugs with different models/components. I don\u2019t see evidence of overclaiming or an unsafe mergeable-pair assertion." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -7220,8 +7197,6 @@ "nodes": [ "issue:24643", "issue:29870", - "issue:34567", - "issue:36331", "issue:38175", "issue:42175", "issue:43257", @@ -7230,6 +7205,7 @@ "issue:43540", "issue:43577", "issue:43582", + "issue:43697", "issue:43716", "issue:43742", "issue:43756", @@ -7264,10 +7240,10 @@ "issue:46693", "issue:46701", "issue:46821", + "issue:46829", "issue:46846" ], "soft_pairs": [ - "issue:34567|issue:36331", "issue:44792|issue:44964", "issue:38175|issue:44898", "issue:42175|issue:44488", @@ -7284,207 +7260,215 @@ "issue:44568|issue:46491", "issue:45405|issue:45600", "issue:45672|issue:45674", - "issue:43906|issue:46468", "issue:45823|issue:46463", - "issue:43531|issue:44821", + "issue:43906|issue:46468", "issue:43577|issue:45072", + "issue:43531|issue:44821", "issue:43906|issue:46693", "issue:44737|issue:46463", "issue:24643|issue:29870", "issue:43452|issue:44488", "issue:44625|issue:44877", - "issue:45663|issue:46846" + "issue:45663|issue:46846", + "issue:43697|issue:46829", + "issue:43540|issue:46693" ], - "prepared_review_unit_hash": "b627bff811932b636b1b4320305de563ad8cdad84ab46376b71a5e43a7658d0f", + "prepared_review_unit_hash": "9c22aef4c3995f45aad7ad369678473bec9327fd319a96411612f51c67925c33", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 27, - "serialized_chars": 35541, - "estimated_input_tokens": 8886, - "estimated_eval_tokens": 18028 + "soft_pair_count": 28, + "serialized_chars": 35882, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: trainer API regressions, tokenizer bugs, multimodal/model-loading failures, quantization issues, and distributed-training problems are all mixed together. There is no clear duplicate core, so all soft edges should be rejected.", - "confidence": 0.98, - "canonical_issue_reason": "No single issue is a clear canonical representative; the items span unrelated subsystems and failure modes rather than one underlying bug.", + "summary": "This cluster is mostly a grab bag of unrelated bug reports spanning model loading, tokenizer/config edge cases, multimodal/video regressions, and quantization/runtime errors. I don\u2019t see any soft pair that clearly describes the same concrete underlying bug or fix.", + "confidence": 0.94, + "canonical_issue_reason": "issue:44479 is the strongest representative: it describes a concrete regression, names multiple affected models, and has the clearest scope/version framing among the set.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a strong global best fit for the cluster because there is no shared underlying problem to anchor on.", + "best_issue_reason": "issue:44479 is the best single issue to keep as the cluster representative because it is specific, user-facing, and documents a broader regression pattern than most others.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:34567", - "right": "issue:36331", - "accept": false, - "reason": "Both are Trainer-related, but one is a state counter bug and the other is a custom compute_loss signature incompatibility; different root causes." - }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Both involve multimodal models, but they affect different models and different failures: a janus test regression vs Phi-4-multimodal loading." + "reason": "Different models and failures: a janus image-generation test issue vs a Phi-4 multimodal load failure." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Different models and symptoms: SigLIP2 zero probabilities vs Perceiver classification with interpolate_pos_encoding." + "reason": "Both are vision-model bugs, but the symptoms and code paths differ: zero probabilities in SigLIP2 vs Perceiver resolution/interpolation failure." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "One is an install/backends packaging issue; the other is a specific model load regression. Not the same bug." + "reason": "Unrelated: packaging/backend dependency omission vs a specific model loading failure for cjvt/sleng-bert." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Different model families and different defects: DeepSpeed weight conversion vs Smollm3 RoPE layer loss." + "reason": "Different underlying problems: DeepSpeed/accelerate weight conversion vs Smollm3 RoPE layer dropping." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Both are video-input regressions, but the affected models and likely code paths differ; not clearly one underlying bug." + "reason": "Similar video-input theme, but the affected model families and reported breakage differ enough that they do not look like the same bug." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Both are tokenizer issues, but the errors are unrelated: AddedToken kwargs conflict vs missing backend_tokenizer." + "reason": "Tokenizer-related, but different failures: AddedToken argument duplication vs Mistral regex patch hitting a missing backend_tokenizer attribute." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Completely different subsystems: init_empty_weights handling vs AutoImageProcessor URL loading." + "reason": "Different loading surfaces and errors: init_empty_weights/_is_hf_initialized vs AutoImageProcessor-from-URL handling." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "Different problems: PIL image processors requiring torchvision vs a Mistral regex patch crash." + "reason": "One is an image processor backend dependency regression; the other is a tokenizer regex patch crash." }, { "left": "issue:43756", "right": "issue:46399", "accept": false, - "reason": "Different models and different mechanisms: RoPE layer dropping vs apply_rotary_pos_emb registration failure." + "reason": "Both touch Qwen-related model internals, but one is RoPE layer selection and the other is kernel registration/decorator handling." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Both mention RoPE/position handling, but they are different models and distinct bugs." + "reason": "Different Qwen-family rotary-position bugs affecting different models and logic paths." }, { "left": "issue:43582", "right": "issue:46821", "accept": false, - "reason": "Unrelated failures: Apple Silicon warmup TypeError vs TimesFM window_size AttributeError." + "reason": "Completely different subsystems: Apple Silicon allocator warmup vs TimesFM window_size attribute handling." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "Different loading failures with different root causes; no shared code-path evidence." + "reason": "Unrelated load-time issues: config key error for MobileLLM vs missing Float8 storage type." }, { "left": "issue:43716", "right": "issue:46701", "accept": false, - "reason": "Both are multimodal issues, but they concern different models and different breakages." + "reason": "Both are multimodal, but one is dtype mismatch in Mistral-3 preprocessing and the other is a DiffusionGemma pipeline failure." }, { "left": "issue:44568", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but one drops BOS/EOS insertion while the other strips a leading space on round-trip." + "reason": "Both are tokenizer regressions, but they involve different tokenizers and different behaviors (special tokens vs encode-decode spacing)." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "Packaging/version pinning issue vs stale auto_mappings config reference; unrelated." + "reason": "Unrelated maintenance issues: PEFT version pin vs stale auto_mappings config references." }, { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Both are 4-bit inference problems, but one is a torch.finfo uint8 issue and the other is a skip-module exclusion logic bug." - }, - { - "left": "issue:43906", - "right": "issue:46468", - "accept": false, - "reason": "No overlap beyond both being model-related issues; they target different reproductions and examples." + "reason": "Both concern 4-bit inference, but the concrete bugs differ: uint8/torch.finfo in Gemma4 audio modules vs llm_int8_skip_modules clearing lm_head exclusion." }, { "left": "issue:45823", "right": "issue:46463", "accept": false, - "reason": "Different areas entirely: Gemma4 device placement vs T5 decoder input validation error messaging." + "reason": "Different problems entirely: Gemma4 device placement vs T5 decoder-input validation messaging." }, { - "left": "issue:43531", - "right": "issue:44821", + "left": "issue:43906", + "right": "issue:46468", "accept": false, - "reason": "Qwen3-MoE sliding_window bug is unrelated to AutoImageProcessor URL loading." + "reason": "No shared bug: one is a reproduction of an existing issue, the other is an official Nemotron example breakage." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Both are dtype-related, but different models and different inference paths; not the same concrete bug." + "reason": "Both mention dtype, but one is Blip2 dtype propagation and the other is bfloat16 inference mismatches in different models." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Different model families and failures: Qwen3-MoE sliding window vs AutoImageProcessor URL loading." }, { "left": "issue:43906", "right": "issue:46693", "accept": false, - "reason": "An isolated reproduction issue is not the same as a long-context FA2 performance regression." + "reason": "A reproduction of an old issue is not the same as a new FA2 long-context performance regression." }, { "left": "issue:44737", "right": "issue:46463", "accept": false, - "reason": "XLNet CPU placement bug and T5 decoder validation message are unrelated." + "reason": "XLNet CPU device-placement bug vs T5 validation error message; no shared code-path." }, { "left": "issue:24643", "right": "issue:29870", "accept": false, - "reason": "Different failures: DeepSpeed training weight-shape RuntimeError vs Bart generate decoder_attention_mask initialization bug." + "reason": "Different training-time failures: DeepSpeed weight shape error vs Bart generate decoder_attention_mask initialization bug." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "Both are loading-related, but one is a gguf/auto model-tokenizer breakage and the other is a specific model load failure; not the same bug." + "reason": "Both are loading issues, but one is a gguf/from_pretrained integration break and the other is a specific model load failure." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Different config-loading problems affecting different models; no shared underlying defect." + "reason": "Different config problems: Qwen3.5 num_labels propagation vs strict config blocking granite_speech loading." }, { "left": "issue:45663", "right": "issue:46846", "accept": false, - "reason": "Different distributed-training bugs on different model families and different mechanisms." + "reason": "Both are Gemma4-related, but one is FSDP2 shared_kv_states rebuilding and the other is missing tensor-parallel plan entries." + }, + { + "left": "issue:43697", + "right": "issue:46829", + "accept": false, + "reason": "Different output-shape/output-field bugs in different models; not the same concrete issue." + }, + { + "left": "issue:43540", + "right": "issue:46693", + "accept": false, + "reason": "One is a video-input ValueError for a Qwen multimodal model, the other a long-context FA2 performance regression; same broad area, different bug." } ] }, @@ -7509,14 +7493,15 @@ "cluster_id": "cluster-10000-502", "nodes": [ "issue:28282", + "issue:29159", "issue:32090", + "issue:33357", "issue:38175", "issue:41762", "issue:43295", "issue:43452", "issue:43525", "issue:43526", - "issue:43540", "issue:43643", "issue:43653", "issue:43688", @@ -7550,7 +7535,6 @@ "issue:46121", "issue:46170", "issue:46463", - "issue:46693", "issue:46701", "issue:46829", "issue:46832", @@ -7558,8 +7542,6 @@ "issue:46932" ], "soft_pairs": [ - "issue:43697|issue:46829", - "issue:43540|issue:46693", "issue:45357|issue:45920", "issue:44913|issue:46121", "issue:28282|issue:32090", @@ -7578,214 +7560,206 @@ "issue:43688|issue:43749", "issue:43452|issue:44843", "issue:43653|issue:44361", - "issue:38175|issue:43295", "issue:43819|issue:44811", + "issue:38175|issue:43295", "issue:43994|issue:44898", "issue:45372|issue:46170", "issue:46832|issue:46932", "issue:43525|issue:46701", - "issue:43295|issue:45356" + "issue:43295|issue:45356", + "issue:29159|issue:33357" ], - "prepared_review_unit_hash": "64bd7454a10f200eb5a4b1d36c6838596096cc4ca917a7fc90e37296a775a9d5", + "prepared_review_unit_hash": "be6826e1cf4925f85952f7cf147fcf659ef6c323d8286b0a8bbb8bf99d76b0ef", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 36474, - "estimated_input_tokens": 9119, - "estimated_eval_tokens": 18494 + "soft_pair_count": 26, + "serialized_chars": 36196, + "estimated_input_tokens": 9049, + "estimated_eval_tokens": 18354 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated bug reports from different model families/subsystems. The only strong duplicate-like pair is the tied-weights serialization regression pair (#44466 / #45005); the rest should stay separate.", - "confidence": 0.87, - "canonical_issue_reason": "#44466 is the clearest, most concrete report of the tied-weights serialization regression (device-dependent lm_head.weight behavior in v5/main), while #45005 is a broader description of the same underlying issue.", + "summary": "This cluster is heterogeneous: most items are distinct model-, tokenizer-, or loading-path bugs that only look loosely similar by surface terms. I rejected all soft-edge pairs as non-duplicates. The most broadly representative issue is the multi-model AutoTokenizer regression, but even that does not justify merging the cluster as duplicates.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45920 is the broadest and most central bug report in the set: it describes an AutoTokenizer regression affecting multiple model families, with a detailed reproduction and wider impact than the more model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "#44466 has the strongest bug report quality for canonicalization: specific regression framing, concrete symptom, and a focused code-path around tied-weight serialization.", + "best_issue_reason": "issue:45920 is the best single issue to represent the cluster because it is the most general, most widely scoped, and best-documented failure among otherwise unrelated edge-case reports.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43697", - "right": "issue:46829", - "accept": false, - "reason": "Different models and different failure modes: RT-DETR output drift vs VideoPrism hidden_states mis-population." - }, - { - "left": "issue:43540", - "right": "issue:46693", - "accept": false, - "reason": "Unrelated problems: Qwen3OmniMoe video input validation vs a FlashAttention performance regression." - }, { "left": "issue:45357", "right": "issue:45920", "accept": false, - "reason": "Both involve model/tokenizer regressions, but they target different behaviors: save_pretrained key serialization vs wrong token IDs during AutoTokenizer load." + "reason": "Both are regression-style loading/serialization bugs, but one is specific to Qwen3.5 visual encoder keys while the other is a broad tokenizer output-ID regression across many models; different code paths and failure modes." }, { "left": "issue:44913", "right": "issue:46121", "accept": false, - "reason": "Both mention RoPE-related config handling, but one is silent reload/default fallback and the other is a TypeError in validation; not the same bug." + "reason": "Both involve RoPE-related config handling, but one is a silent config reload defaulting issue and the other is a type error in JSON-loaded lists during validation; not the same underlying bug." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "Completely different issues: missing PyTorch import vs a Trainer GPU broadcast TypeError." + "reason": "An import-time PyTorch missing dependency error is unrelated to a Trainer GPU broadcast NoneType error; different subsystems and no shared code-path bug." }, { "left": "issue:46463", "right": "issue:46899", "accept": false, - "reason": "Different subsystems and symptoms: T5 decoder-input validation messaging vs Gemma4 quantized multimodal embedding dtype casting." + "reason": "The T5 validation error-message bug and the Gemma4 quantized embedders dtype-cast bug are unrelated; one is messaging/validation, the other is a runtime quantization issue." }, { "left": "issue:44737", "right": "issue:45588", "accept": false, - "reason": "XLNet CPU device placement bug and flash_attention s_aux=None AttributeError are unrelated." + "reason": "XLNet CPU arange placement and flash-attention s_aux=None handling are separate implementation bugs in different modules." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "Different regressions in different models: RT-DETR output changes vs Qwen3.5 save_pretrained checkpoint corruption." + "reason": "RT-DETR output mismatch and Qwen3.5 incorrect saved checkpoint keys are different regressions affecting different save/load or forward paths." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to the single-label-classification zero-loss problem." + "reason": "BEiTImageProcessorFast returning only one reduced label is unrelated to the zero-loss classification problem across sequence-classification models." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "BitNet accelerate-loading unpacking bug is not the same as Perceiver interpolation/flexible-resolution failure." + "reason": "BitNet accelerated-loading unpacking and Perceiver interpolation failures are distinct model-specific bugs with no shared concrete code-path." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Both are config-field issues, but one is AutoConfig with trust_remote_code and the other is a model-specific strict-field omission; not the same defect." + "reason": "One is a generic AutoConfig trust_remote_code field-loss issue; the other is a specific Qwen3_5MoeVisionConfig field being dropped by strict serialization. Similar theme, not the same bug." }, { "left": "issue:44466", "right": "issue:45005", - "accept": true, - "reason": "Same underlying tied-weights serialization regression in v5: both describe incorrect handling of tied lm_head/translation-model weights during save/load." + "accept": false, + "reason": "Both mention tied weights, but one is device-dependent lm_head serialization and the other is translation-model tied-weight issues; too broad and not clearly one fixable code-path." }, { "left": "issue:45200", "right": "issue:46463", "accept": false, - "reason": "Gemma4 text-only mm_token_type_ids defaulting vs T5 decoder validation messaging are unrelated." + "reason": "Gemma 4 missing mm_token_type_ids defaults and T5 decoder-input validation error messaging are unrelated bugs." }, { "left": "issue:44493", "right": "issue:46050", "accept": false, - "reason": "Unexpected position-id keys and sinq quantization save/load breakage are different serialization issues." + "reason": "Unexpected position-id keys and sinq quantization save/load breakage are separate serialization problems." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "DeepSpeed ZeRO-3 loading index error and trainer.evaluate() after train are different training/runtime failures." + "reason": "DeepSpeed ZeRO-3 loading crash and trainer.evaluate() after train are different training/runtime failures." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "Different tokenizers and different failure points: MLukeTokenizer task AttributeError vs AutoTokenizer loading FastSpeech2ConformerTokenizer." + "reason": "Both are tokenizer-related, but one is an AttributeError in MLukeTokenizer on tasks and the other is AutoTokenizer failing to load a specific tokenizer class; not the same underlying defect." }, { "left": "issue:43761", "right": "issue:46829", "accept": false, - "reason": "Both concern hidden_states, but one returns None despite request and the other mislabels outputs in a different model; not the same code-path bug." + "reason": "Both concern hidden_states in outputs, but one omits hidden_states entirely while the other misplaces last_hidden_state into hidden_states; different models and output paths." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Auxiliary-loss normalization in MoE models is unrelated to FSDP CPU RAM efficient loading." + "reason": "Aux-loss normalization in MoE models is unrelated to FSDP CPU RAM efficient loading." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "Both mention AutoTokenizer/from_pretrained, but gguf_file support breakage and unconditional model_info() in offline mode are distinct issues." + "reason": "gguf_file loading breakage and HF_HUB_OFFLINE model_info calls in mistral regex patch are separate tokenizer/model-loading regressions." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBird mask-token registration/decoding bug is unrelated to MLukeTokenizer task AttributeError." + "reason": "BigBird special-token decode behavior and MLukeTokenizer task AttributeError are both tokenizer issues but not the same failure mode or code path." }, { - "left": "issue:38175", - "right": "issue:43295", + "left": "issue:43819", + "right": "issue:44811", "accept": false, - "reason": "Siglip2 zero-probabilities regression is unrelated to processor.tokenizer access / image-to-tokenizer regression in v4.57.5." + "reason": "DAC latent conversion mismatch and Whisper batch_decode skip_special_tokens handling are unrelated model vs processor bugs." }, { - "left": "issue:43819", - "right": "issue:44811", + "left": "issue:38175", + "right": "issue:43295", "accept": false, - "reason": "DAC.from_latents vs forward mismatch is a model correctness bug; Whisper batch_decode skip_special_tokens handling is a tokenizer/processor API bug." + "reason": "Siglip2 zero probabilities and processor.tokenizer regression are distinct multimodal issues with different symptoms and likely fixes." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Different models and failures: SigLIP2 nonsensical outputs vs Perceiver non-default-resolution interpolation failure." + "reason": "Siglip2 nonsensical outputs and Perceiver non-default resolution failure are unrelated model-specific inference bugs." }, { "left": "issue:45372", "right": "issue:46170", "accept": false, - "reason": "Gemma 4 processor dependency import failure and Windows ROCm torch.distributed import/type issue are unrelated environment problems." + "reason": "A missing mistral_common import for Gemma4 processor loading is unrelated to a Windows ROCm torch.distributed dependency issue." }, { "left": "issue:46832", "right": "issue:46932", "accept": false, - "reason": "RT-DETR feature-level mismatch and Gemma4 quantized embedders dtype-cast issue are unrelated." + "reason": "RT-DETR feature-level mismatch and Gemma4 quantized embedder dtype casting are completely different model paths and fixes." }, { "left": "issue:43525", "right": "issue:46701", "accept": false, - "reason": "Llama4Config pad_token_id attribute error is unrelated to DiffusionGemma image-text-to-text pipeline failure." + "reason": "Llama4Config missing pad_token_id and DiffusionGemmaGenerationOutput pipeline failure are unrelated config vs pipeline issues." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Both are regression reports, but one is about processor/tokenizer access and image passing, while the other is Kimi tokenizer codec/regex handling." + "reason": "Both are tokenizer regressions, but one is about processor.tokenizer access/images passing and the other is codec handling plus a misleading warning; not one bug." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency is unrelated to a MacOS bus error in a CLIP community model." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The analysis is well grounded in the packet and stays conservative on the soft-edge pairs. The only duplicate-like match that is clearly supported by the titles is issue:44466 / issue:45005 around tied-weights serialization in v5. The rest of the pairwise rejections are appropriately cautious and do not overclaim sameness from superficial similarity." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -7802,9 +7776,7 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-502", "nodes": [ - "issue:29159", "issue:32090", - "issue:33357", "issue:33666", "issue:36010", "issue:41720", @@ -7814,6 +7786,7 @@ "issue:43531", "issue:43550", "issue:43577", + "issue:43611", "issue:43646", "issue:43653", "issue:43688", @@ -7848,14 +7821,14 @@ "issue:46491", "issue:46612", "issue:46693", - "issue:46821" + "issue:46821", + "issue:46846" ], "soft_pairs": [ - "issue:29159|issue:33357", "issue:43653|issue:44568", "issue:43646|issue:43749", - "issue:45542|issue:46178", "issue:43531|issue:43742", + "issue:45542|issue:46178", "issue:42898|issue:46491", "issue:43688|issue:45704", "issue:43756|issue:46402", @@ -7875,207 +7848,208 @@ "issue:44734|issue:46612", "issue:44977|issue:46693", "issue:43504|issue:43550", - "issue:43504|issue:43720", "issue:44464|issue:44977", + "issue:43504|issue:43720", "issue:41720|issue:44464", - "issue:43577|issue:44554" + "issue:43577|issue:44554", + "issue:43611|issue:46846" ], - "prepared_review_unit_hash": "4cae980c9272dd5e419b5cf95e7c043f009fe650f5d5654f018ed98310b9aad2", + "prepared_review_unit_hash": "3b4e64e934e3281bb6c86be9d8ab727f390e4ae0d597e816834176a87b7fcbe1", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35560, - "estimated_input_tokens": 8890, - "estimated_eval_tokens": 18036 + "serialized_chars": 35594, + "estimated_input_tokens": 8899, + "estimated_eval_tokens": 18054 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are a loose mix of unrelated transformer bugs/regressions; none of the soft pairs look like safe duplicates, so no pair is accepted.", - "confidence": 0.78, - "canonical_issue_reason": "No single canonical issue is appropriate: the set spans different subsystems (tokenizers, generation, loading, quantization, backend availability, model-specific regressions) and the paired items do not collapse to one underlying bug.", + "summary": "The cluster is heterogeneous: most items are unrelated Transformers regressions across tokenization, loading, generation, model configs, and backend/device-specific behavior. None of the soft pairs look like true duplicates or the same fixable code-path bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43646 is the broadest representative of the cluster because it concerns a generic Transformers v5 model initialization breakage, but it is not a true duplicate target for most other items.", "canonical_pr_reason": null, - "best_issue_reason": "If one issue had to stand in as a representative, #29159 is the broadest user-facing report, but it still only covers a tokenizer inconsistency and does not represent the rest of the cluster well.", + "best_issue_reason": "issue:43646 is the best single issue to represent the cluster if one must be chosen, since it is high-level and v5-related; however, the surrounding issues are too diverse for strong canonicalization.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:29159", - "right": "issue:33357", - "accept": false, - "reason": "Tokenizer inconsistency vs MacOS CLIP bus error; clearly different failure modes." - }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Both concern special-token handling, but they affect different tokenizers and different behavior (mask token decode vs BOS/EOS insertion)." + "reason": "Both are tokenizer/special-token bugs, but one is BigBird mask-token decoding and the other is mDeBERTa BOS/EOS handling; different failure modes." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model init breakage and FSDP CPU RAM efficient loading are unrelated code paths." + "reason": "Custom model initialization in v5 is unrelated to FSDP CPU RAM efficient loading; different code paths and symptoms." }, { - "left": "issue:45542", - "right": "issue:46178", + "left": "issue:43531", + "right": "issue:43742", "accept": false, - "reason": "Both mention backend availability, but one is tensorboard/TensorFlow absence and the other is an invalid tensorflow_text backend name." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to a MobileLLM key error while loading pretrained weights." }, { - "left": "issue:43531", - "right": "issue:43742", + "left": "issue:45542", + "right": "issue:46178", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key-error loading are unrelated." + "reason": "Both mention backend setup, but one is undefined tf backend after installing tensorboard and the other is a missing BACKENDS_MAPPING entry for tensorflow_text." }, { "left": "issue:42898", "right": "issue:46491", "accept": false, - "reason": "Both are tokenizer regressions, but one is cleanup-space behavior and the other is a leading-space round-trip bug in CodeLlama." + "reason": "One is a v5 clean_up_tokenization_spaces behavior change; the other is a CodeLlama leading-space round-trip regression." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Auxiliary-loss normalization and an apex RMSNorm memory leak are different bugs." + "reason": "Aux-loss normalization in MoE models is unrelated to T5's apex.FusedRMSNorm memory leak." }, { "left": "issue:43756", "right": "issue:46402", "accept": false, - "reason": "Smollm3 RoPE-layer loading and Sam3 text-encoder weight loading are unrelated model-specific issues." + "reason": "Smollm3 RoPE-layer selection and Sam3 text-encoder weight loading are different model-specific bugs." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Both involve tokenizer loading, but the offline hub lookup bug and the est-roberta tokenizer regression are different causes." + "reason": "Offline model_info() misuse in _patch_mistral_regex is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "GGUF loading failures and Blip2 dtype propagation are unrelated." + "reason": "gguf_file loading failures and BLIP2 dtype propagation are distinct issues with different APIs and errors." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Special-token registration and DiaConfig token-id persistence are distinct serialization/tokenizer issues." + "reason": "BigBird special-token registration and DiaConfig losing custom token IDs after save/load are different tokenizer/config persistence bugs." }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Chunked generation with compiled forward and Qwen3.5 recurrent-state reset are different cache/generation bugs." + "reason": "Both affect generation with cache, but one is chunked generation under compiled forward and the other is recurrent states resetting in modular_qwen3_5." }, { "left": "issue:43688", - "right": "issue:45440", + "right": "issue:45701", "accept": false, - "reason": "Aux-loss normalization and DeepseekV3MoE divergence do not describe the same change or bug." + "reason": "Aux-loss normalization is unrelated to version-dependent tokenization changes." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video input regression and num_labels propagation are unrelated Qwen3.5 issues." + "reason": "Video-input regression on Qwen VL models is unrelated to num_labels propagation in Qwen3.5 configs." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "Both mention NaNs, but one is Gemma3 mixed-length sliding-window embeddings and the other is classification-head NaNs after checkout/init." + "reason": "Both mention NaNs, but one is Gemma3 sliding-window attention on GPU and the other is classification-head init/checkout behavior." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "A save_pretrained shard-size question and AutoProcessor kwargs forwarding are unrelated." + "reason": "A question about save_pretrained shard size is unrelated to AutoProcessor failing to forward kwargs to cached_file." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Different model families and different symptoms; no shared underlying bug." + "reason": "MoE aux-loss normalization and DeepseekV3MoE parity with remote implementation are not the same concrete bug." }, { "left": "issue:44821", "right": "issue:46821", "accept": false, - "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated." + "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated loader/API issues." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Multi-GPU training for Qwen2-VL and a GenerationMixin import error are unrelated." + "reason": "Qwen2-VL multi-GPU training and the GenerationMixin import error are unrelated." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer GPU broadcast NoneType error vs Qwen2-VL multi-GPU training are different problems." + "reason": "Trainer _gpu_broadcast_one NoneType error is unrelated to Qwen2-VL multi-GPU training." }, { "left": "issue:44466", "right": "issue:46050", "accept": false, - "reason": "lm_head serialization inconsistency and sinq quantized save/load breakage are distinct persistence bugs." + "reason": "lm_head.weight serialization inconsistency and sinq quantized save/load breakage involve different model-state persistence problems." }, { "left": "issue:44734", "right": "issue:46612", "accept": false, - "reason": "KV-cache tensor-indexing crash in serve and beam-search cache reorder omission across models are different generation bugs." + "reason": "Wrong tensor indexing in serve KV-cache continuation is different from beam-search cache reorder being skipped for certain models." }, { "left": "issue:44977", "right": "issue:46693", "accept": false, - "reason": "Qwen3.5 flash-attention correctness and FA2 long-context performance regression are not the same bug." + "reason": "Qwen3.5 flash-attention generation failure is distinct from a long-context FA2 performance regression." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Legacy-field loading failure in BeitForSemanticSegmentation and Bamba torch.compile/SDPA failure are unrelated." + "reason": "Beit legacy-field load failure and Bamba torch.compile+SDPA failure are different model/runtime bugs." }, { - "left": "issue:43504", - "right": "issue:43720", + "left": "issue:44464", + "right": "issue:44977", "accept": false, - "reason": "Beit legacy-field loading and BitNet packed-weight unpacking during accelerate loading are different model-loading issues." + "reason": "Chunked generation with compiled forward is not the same bug as Qwen3.5 flash-attention generation failure." }, { - "left": "issue:44464", - "right": "issue:44977", + "left": "issue:43504", + "right": "issue:43720", "accept": false, - "reason": "Compiled-forward chunked generation and Qwen3.5 flash-attention generation bug are different code paths." + "reason": "Beit loading with a legacy field and BitNet packed-weight unpacking during accelerate loading are unrelated." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "MPS attention correctness on A800 and compiled chunked generation are unrelated." + "reason": "Qwen3 auto device-mapping cudaErrorAssert on A800 is a different failure than compiled chunked generation inconsistency." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "Blip2 dtype propagation and an MPS attention correctness issue are not the same underlying bug." + "reason": "BLIP2 dtype propagation and MPS attention correctness with mismatched value/query head dims are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:46846", + "accept": false, + "reason": "base_model_prefix loading breakage in v5 and missing linear_attn entries in the TP plan are different config/loading issues." } ] }, @@ -8114,10 +8088,10 @@ "issue:43525", "issue:43572", "issue:43606", - "issue:43611", "issue:43643", "issue:43646", "issue:43688", + "issue:43701", "issue:43720", "issue:43782", "issue:43825", @@ -8138,6 +8112,7 @@ "issue:45290", "issue:45405", "issue:45584", + "issue:45812", "issue:45823", "issue:46050", "issue:46192", @@ -8145,11 +8120,9 @@ "issue:46365", "issue:46489", "issue:46693", - "issue:46710", - "issue:46846" + "issue:46710" ], "soft_pairs": [ - "issue:43611|issue:46846", "issue:43643|issue:44561", "issue:43782|issue:46693", "issue:43720|issue:43927", @@ -8176,204 +8149,219 @@ "issue:43278|issue:43844", "issue:44568|issue:46489", "issue:46209|issue:46710", - "issue:43646|issue:44466" + "issue:43646|issue:44466", + "issue:44568|issue:45812", + "issue:33357|issue:33666", + "issue:43278|issue:43701" ], - "prepared_review_unit_hash": "629ab75fdc2ff9f36c4369f9c2b28a94fc829c06a1ee9d4b58cefa9380140cce", + "prepared_review_unit_hash": "c5cf00e1ee3f38a47d662aadd8de556ac09d65d50c189f78c49a6010c3986639", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36259, - "estimated_input_tokens": 9065, - "estimated_eval_tokens": 18386 + "soft_pair_count": 30, + "serialized_chars": 36807, + "estimated_input_tokens": 9202, + "estimated_eval_tokens": 18660 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are related only at a very broad level: many are Transformers v5 regressions or tokenizer/model-loading bugs, but they affect different code paths and are not duplicates of one another. I found no soft pair that looks like the same underlying bug or mergeable change.", - "confidence": 0.72, - "canonical_issue_reason": "issue:43646 is the broadest representative of the cluster\u2019s apparent theme: Transformers 5.0 breakages around custom model initialization and related model-loading regressions.", + "summary": "This cluster is a mix of largely unrelated bug reports. The soft pairs mostly share broad vocabulary (tokenizer/config/regression/processor) but describe different models, code paths, and failure modes, so none should be deduplicated.", + "confidence": 0.79, + "canonical_issue_reason": "If forced to choose one representative issue, issue 45812 is the strongest candidate: it is a concrete tokenizer regression affecting Granite models, has multiple inbound references, and describes a clear code-path bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 best fits as the cluster representative because it is general, central to the v5 regression theme, and not narrowly tied to one model family or edge case.", + "best_issue_reason": "Issue 45812 is the best single representative of the cluster\u2019s only recurring theme (tokenizer regressions), with the clearest impact and most actionable report among the listed issues.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43611", - "right": "issue:46846", - "accept": false, - "reason": "Both mention loading issues, but one is a base_model_prefix regression and the other is a Qwen3.5 tensor-parallel planning/OOM problem; different bugs and code paths." - }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Both are trust_remote_code-related regressions, but one is missing fields from AutoConfig and the other is removal of is_torch_fx_available breaking remote models; distinct failures." + "reason": "Both involve v5/trust_remote_code breakage, but one is missing fields from AutoConfig and the other is removal of is_torch_fx_available; different bugs and fixes." }, { "left": "issue:43782", "right": "issue:46693", "accept": false, - "reason": "Qwen3VL from_pretrained weight_only failure is unrelated to a FA2 long-context generation performance regression." + "reason": "Qwen3VL from_pretrained weight_only failure and FA2 long-context performance regression are unrelated code paths." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "One is packed-weight unpacking during accelerate loading; the other is DiaConfig custom token IDs being lost on save/load. Different model subsystems." + "reason": "BitNet packed-weight loading vs DiaConfig token-ID loss on save/load are distinct failures." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Multi-GPU training for Qwen2-VL and a CustomTrainer compute_loss signature change are unrelated training issues." + "reason": "Multi-GPU Qwen2-VL training and a CustomTrainer compute_loss signature mismatch are unrelated." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "These both involve model behavior differences, but one is an embedding dtype regression and the other is auxiliary loss normalization; not the same bug." + "reason": "Embedding dtype drift during evaluate is not the same as auxiliary-loss normalization in MoE models." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "BEiT legacy-field loading failure and OmDet-Turbo processor image-size mismatch affect different model families and different failure modes." + "reason": "Legacy field loading for Beit semantic segmentation and OmDet processor image-size mismatch are different bugs." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch in bark-small is unrelated to DiaConfig token ID persistence and generation index errors." + "reason": "CPU offload device mismatch and DiaConfig custom token-ID corruption are different failure modes." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio processor sampling-rate defaulting and processor.tokenizer/image-passing regression are different processor behaviors." + "reason": "Audio processor sampling-rate defaulting and processor.tokenizer/image handling regression do not share the same underlying defect." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 implementation discrepancy and pipeline error-message support mismatch are unrelated issues." + "reason": "SigLIP2 implementation discrepancy and a pipeline error message regression are unrelated." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "Packed-weight loading failure and MLukeTokenizer task AttributeError are different code paths and failure types." + "reason": "Packed-weight loading in BitNet and MLukeTokenizer AttributeError are distinct issues." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Model loading failure for Phi-4 multimodal and MIN_PEFT_VERSION being bumped too early are unrelated release/load issues." + "reason": "Phi-4 multimodal model loading failure and an unreleased PEFT version bump are unrelated." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Batched Qwen2_5_VLProcessor chat template crash and assistant tool-call content crash are both chat-template related but different edge cases and inputs." + "reason": "Both hit apply_chat_template, but one is batched padding=False on Qwen2.5-VL and the other is assistant tool-call handling; different edge cases." }, { "left": "issue:44554", "right": "issue:46365", "accept": false, - "reason": "MPS attention correctness with differing head dims is unrelated to Trainer import conflicts caused by datasets." + "reason": "MPS attention correctness and importing Trainer failing due to datasets are completely unrelated." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 flash-attention generation and FP8 inference on B200 are different hardware/backend issues." + "reason": "Qwen3.5 flash-attention generation problems and FP8 inference on B200 are different bugs." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Whisper empty transcription after align_special_tokens are unrelated model/config bugs." + "reason": "num_labels propagation in Qwen3.5 and Whisper empty-transcription generation failure are unrelated." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Llama4 missing pad_token_id and StableLm missing pad_token_idx are similar in shape but affect different config fields and model families." + "reason": "Missing pad_token_id in Llama4Config and missing pad_token_idx in StableLmConfig are similar naming, but different configs and different breakages." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "gguf_file breaking AutoTokenizer/AutoModel loading and failure to load vesteinn/ScandiBERT are separate loading problems with no clear shared root cause." + "reason": "gguf_file breaking tokenizer/model loading and a ScandiBERT load failure are not the same defect." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Gemma4 device mismatch under device_map=auto and sinq quantization save/load breakage are unrelated model execution vs serialization issues." + "reason": "Gemma4 device-map mismatch and sinq save/load quantization breakage are unrelated." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer extra-space bug and MacOS bus error on a CLIP model are entirely different problems." + "reason": "Tokenizer spacing after added tokens and a MacOS bus error in CLIP are unrelated." }, { "left": "issue:43994", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 nonsensical outputs and return_tensors being ignored in text_kwargs are different processor/pipeline bugs." + "reason": "SigLIP2 nonsensical outputs and return_tensors being ignored when text_kwargs is passed are different issues." }, { "left": "issue:42796", "right": "issue:44977", "accept": false, - "reason": "Ministral tokenizer BPE-marker handling and Qwen3.5 flash-attention generation are unrelated tokenizer vs attention/backend issues." + "reason": "Ministral BPE marker handling and Qwen3.5 flash-attention generation are unrelated." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 implementation discrepancy and gradient blow-up with HfDeepSpeedConfig + ZeRO-3 are different numerical/optimization issues." + "reason": "A model discrepancy and a DeepSpeed gradient regression do not share the same code-path problem." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 docs example errors and processor.tokenizer/image-passing regression are both processor-related but not the same bug or fix." + "reason": "Documentation/example errors for SigLIP2 are not the same as the processor.tokenizer/images regression." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Embedding dtype drift in evaluate and gradient increase under ZeRO-3 are different training-time regressions." + "reason": "Embedding dtype change and ZeRO-3 gradient growth are unrelated training behaviors." }, { "left": "issue:44568", "right": "issue:46489", "accept": false, - "reason": "mdeberta add_special_tokens BOS/EOS regression and DeepSeek-Coder tokenizer BPE-marker regression are different tokenizer behaviors." + "reason": "Both are tokenizer bugs, but they affect different models and mechanisms (special-token insertion vs DeepSeek-Coder BPE regression)." }, { "left": "issue:46209", "right": "issue:46710", "accept": false, - "reason": "FP8 inference failure on B200 and accuracy regression after upgrading Transformers are unrelated hardware/backend vs model-quality regressions." + "reason": "FP8 inference failure on B200 and an accuracy regression after upgrade are different problems." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization breakage and device-dependent lm_head.weight serialization are separate v5 regressions with different code paths." + "reason": "Custom model initialization breakage in v5 and inconsistent lm_head serialization are distinct issues." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both involve tokenization, but one is add_special_tokens behavior for mdeberta and the other is wrong token IDs for Granite; not the same bug." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "MacOS bus error on CLIP and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Embedding dtype drift in evaluation and resume_from_checkpoint key mismatch are different defects." } ] }, @@ -8397,10 +8385,7 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-502", "nodes": [ - "issue:33357", - "issue:33666", "issue:43257", - "issue:43278", "issue:43299", "issue:43334", "issue:43504", @@ -8408,23 +8393,23 @@ "issue:43575", "issue:43606", "issue:43653", - "issue:43701", "issue:43824", "issue:43844", "issue:43873", "issue:43927", "issue:43937", + "issue:43994", "issue:44112", "issue:44291", "issue:44361", "issue:44464", "issue:44488", "issue:44493", - "issue:44568", "issue:44589", "issue:44792", "issue:44843", "issue:44849", + "issue:44869", "issue:44964", "issue:45072", "issue:45200", @@ -8433,8 +8418,9 @@ "issue:45357", "issue:45405", "issue:45520", + "issue:45588", "issue:45759", - "issue:45812", + "issue:45823", "issue:46050", "issue:46209", "issue:46285", @@ -8442,25 +8428,23 @@ "issue:46424", "issue:46701", "issue:46710", + "issue:46752", "issue:46821", "issue:46832", "issue:46846" ], "soft_pairs": [ - "issue:44568|issue:45812", - "issue:33357|issue:33666", "issue:44488|issue:44843", - "issue:43278|issue:43701", "issue:44849|issue:46209", "issue:44792|issue:45405", "issue:43927|issue:44361", "issue:43575|issue:45405", "issue:43504|issue:43606", - "issue:43824|issue:44589", "issue:44849|issue:45357", + "issue:43824|issue:44589", "issue:43257|issue:43299", - "issue:44493|issue:46285", "issue:45245|issue:46710", + "issue:44493|issue:46285", "issue:46399|issue:46846", "issue:46821|issue:46832", "issue:43606|issue:44112", @@ -8474,211 +8458,224 @@ "issue:43937|issue:46701", "issue:43844|issue:44964", "issue:43873|issue:46050", - "issue:45200|issue:45229" + "issue:45200|issue:45229", + "issue:44869|issue:46752", + "issue:45229|issue:45823", + "issue:43994|issue:44493", + "issue:45588|issue:46050" ], - "prepared_review_unit_hash": "8da0f9bd71b3e4b03aa37679f8ed86774a3f68afa02177098dc3cc2f4cd3fabb", + "prepared_review_unit_hash": "fa7b28f5672b2c5a5090712f76ed79be6f5569b74aa09bd2c7044fe2cf5e8586", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36153, - "estimated_input_tokens": 9039, - "estimated_eval_tokens": 18334 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35959, + "estimated_input_tokens": 8990, + "estimated_eval_tokens": 18236 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly unrelated transformer/model-loading bugs, so no soft pair should be merged; I\u2019d use the broad Granite tokenizer regression as the closest representative issue. No PRs were provided.", - "confidence": 0.78, - "canonical_issue_reason": "issue:45812 is the broadest, most clearly regression-shaped issue in the set: it describes a silent v4\u2192v5 tokenizer break affecting all Granite models and has the strongest external impact signal among the issues.", + "summary": "Mostly heterogeneous. The clearest subcluster is Qwen3 MoE loading regressions, with 43299 as the broadest representative. Qwen multimodal/config-loading reports are related by theme but not strong enough to treat as the same bug, especially 43334 vs 45759.", + "confidence": 0.31, + "canonical_issue_reason": "issue:43299 is the broadest early Qwen3/Qwen3VL MoE loading regression in the packet, so it best represents the small loading-failure subcluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the best single issue to represent the cluster because it is concrete, version-linked, and broadly applicable; the other issues are mostly model-specific or unrelated failure modes.", + "best_issue_reason": "issue:43299 is the best single anchor for the cluster\u2019s only clearly related issue group; the rest of the packet is too heterogeneous to justify a broader canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44568", - "right": "issue:45812", - "accept": false, - "reason": "Different models and symptoms: mdeberta-v3 special-token handling vs Granite tokenizer ID regression." - }, - { - "left": "issue:33357", - "right": "issue:33666", - "accept": false, - "reason": "Unrelated problems: MacOS CLIP bus error vs Qwen2-VL multi-GPU training." - }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "Both involve tokenizer/model loading, but one is a specific model load failure and the other is an offline-mode regression in a Mistral regex patch." - }, - { - "left": "issue:43278", - "right": "issue:43701", - "accept": false, - "reason": "Embedding dtype mismatch during eval is unrelated to resume_from_checkpoint key mismatches." + "reason": "Both are loading-related, but one is a specific model-load failure and the other is an offline-mode tokenizer patch regression; different code paths." }, { "left": "issue:44849", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 hidden-state output bug is a different code path from FP8 inference failure on B200." + "reason": "Different model families and failure modes: Qwen3.5 hidden-state behavior vs FP8 inference on B200." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "A Janus test failure and an unmet PEFT version requirement are unrelated." + "reason": "Janus generation test failure vs a PEFT version packaging issue; not the same bug." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "Saving/loading custom token IDs causing generation IndexError is different from MLuke tokenizer task-time AttributeError." + "reason": "Distinct tokenizer/config problems with different symptoms and affected models." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "Tensor-parallel OOM for Qwen2-57B is unrelated to the PEFT version bump issue." + "reason": "Tensor-parallel OOM during loading is unrelated to the PEFT version bump issue." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field loading failure in BEiT is unrelated to Bark CPU-offload device mismatch." + "reason": "Legacy-field model loading and CPU offload device mismatch are different bugs." }, { - "left": "issue:43824", - "right": "issue:44589", + "left": "issue:44849", + "right": "issue:45357", "accept": false, - "reason": "Importing a Qwen2.5-VL class and missing Float8 storage are different failures." + "reason": "Same model family, but hidden-states behavior and save_pretrained key regression are different concrete issues." }, { - "left": "issue:44849", - "right": "issue:45357", + "left": "issue:43824", + "right": "issue:44589", "accept": false, - "reason": "Hidden-states behavior in Qwen3.5 is unrelated to incorrect visual encoder keys on save_pretrained." + "reason": "Import error for a Qwen class and Float8 storage lookup failure are unrelated compatibility problems." }, { "left": "issue:43257", "right": "issue:43299", - "accept": false, - "reason": "Both concern Qwen MOE loading, but one is accelerate/deepspeed weight conversion and the other is a transformers-dev model-loading breakage." + "accept": true, + "reason": "Both are Qwen3/Qwen3-VL MoE loading regressions on the dev/5.0 line; a shared fix in the loading/conversion path could plausibly cover both." }, { - "left": "issue:44493", - "right": "issue:46285", + "left": "issue:45245", + "right": "issue:46710", "accept": false, - "reason": "Unexpected position-id keys across many models is not the same as the CLIP text-model structure change." + "reason": "Category-count runtime error and an accuracy regression are unrelated." }, { - "left": "issue:45245", - "right": "issue:46710", + "left": "issue:44493", + "right": "issue:46285", "accept": false, - "reason": "Category cardinality error is a distinct bug from an accuracy regression after upgrade." + "reason": "Unexpected position-id keys are unrelated to the CLIP text-model structural change." }, { "left": "issue:46399", "right": "issue:46846", "accept": false, - "reason": "Missing decorator for apply_rotary_pos_emb is a different issue than missing linear_attn TP-plan entries causing OOM/shape errors." + "reason": "Both are Qwen3.5-related, but one is a kernel registration error and the other is a tensor-parallel plan/OOM issue." }, { "left": "issue:46821", "right": "issue:46832", "accept": false, - "reason": "TimesFM window_size attribute access and RT-DETR feature-level mismatch are unrelated model-specific bugs." + "reason": "TimesFM window_size AttributeError and RT-DETR feature-level mismatch are unrelated." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "CPU offload device mismatch is not the same as a stale GraniteSpeech CI test." + "reason": "The CI test failure is not the same bug as the Bark CPU-offload device mismatch." }, { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Both touch Qwen multimodal configs, but one is a missing pad_token_id attribute and the other is a text_config unwrapping issue." + "reason": "Related Qwen multimodal/config-loading area, but the packet only shows distinct titles and symptoms; not enough evidence they are the same underlying bug." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Qwen3-VL loading failure is unrelated to StableLmConfig missing pad_token_idx." + "reason": "Both are config-field regressions after 5.0, but they affect different models and different missing/changed fields." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "BEiT legacy-field loading and BigBirdTokenizer special-token registration are different bugs." + "reason": "Legacy-field model loading and tokenizer special-token registration are unrelated." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Missing Float8 storage and a Python 3.13 flash_attn KeyError are separate compatibility issues." + "reason": "Float8 storage lookup and flash_attn import on Python 3.13 are different compatibility issues." }, { "left": "issue:45072", "right": "issue:46832", "accept": false, - "reason": "Bfloat16 dtype mismatches in inference are unrelated to RT-DETR forward feature-count errors." + "reason": "Dtype mismatch in inference and RT-DETR feature-level indexing/type error are different subsystems." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and unexpected _is_hf_initialized in init_empty_weights are different version-compatibility bugs." + "reason": "Missing config attribute and unexpected init argument are related only at a high level, not the same fix." }, { "left": "issue:44464", "right": "issue:46424", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is not the same as a cache-length limitation for quantized cache." + "reason": "Compiled-forward chunked generation inconsistency is unrelated to cache max length handling." }, { "left": "issue:43937", "right": "issue:46701", "accept": false, - "reason": "Invalid GenerationConfig in GLM-5 is unrelated to DiffusionGemmaGenerationOutput pipeline failure." + "reason": "GenerationConfig validation and DiffusionGemma pipeline output issues are unrelated." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "ZeRO-3 gradient growth and Phi-4 multimodal loading failure are unrelated." + "reason": "ZeRO-3 gradient instability and a multimodal model load failure are unrelated." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Quantization offloading behavior differs from save/load breakage in sinq-quantized models." + "reason": "Quantization/offloading and sinq save/load breakage are different code paths." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Text-only fine-tuning defaults for Gemma 4 are unrelated to Gemma4 multi-GPU inference OOM." + "reason": "Text-only token defaults and multi-GPU CUDA OOM are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:46752", + "accept": false, + "reason": "Whisper timestamp decoding crash and empty chat-template IndexError are different edge cases." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma4 inference issues, but one is OOM and the other is a device mismatch; not the same bug." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 incorrect outputs and unexpected position-id keys are unrelated." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention sink-less-model crash is unrelated to sinq serialization/deserialization breakage." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, + "evaluator_result": { + "accept": false, + "feedback": "Mostly grounded, but one soft-edge acceptance is too strong for the evidence available. In particular, issue:43334 vs issue:45759 looks related at the Qwen multimodal/config level, but the packet only shows distinct titles ('missing pad_token_id' vs 'does not unwrap text_config') and does not justify treating them as likely the same underlying bug. That makes the summary's 'plausible duplicate subclusters' claim a bit overstated. The other soft-edge rejections look conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -8698,6 +8695,7 @@ "issue:15354", "issue:28218", "issue:29159", + "issue:29870", "issue:29942", "issue:30333", "issue:32090", @@ -8712,46 +8710,41 @@ "issue:43450", "issue:43493", "issue:43540", + "issue:43550", + "issue:43572", "issue:43696", "issue:43701", + "issue:43844", "issue:43873", "issue:43957", - "issue:43994", "issue:44186", "issue:44423", - "issue:44493", "issue:44554", "issue:44610", "issue:44734", - "issue:44869", "issue:44977", "issue:45072", "issue:45084", "issue:45198", "issue:45200", - "issue:45229", "issue:45406", "issue:45446", "issue:45520", "issue:45588", "issue:45759", "issue:45823", - "issue:46050", "issue:46121", + "issue:46178", "issue:46192", "issue:46369", "issue:46701", "issue:46710", - "issue:46752" + "issue:46821" ], "soft_pairs": [ - "issue:45229|issue:45823", - "issue:44869|issue:46752", - "issue:43994|issue:44493", - "issue:45588|issue:46050", "issue:29942|issue:30333", - "issue:29159|issue:32090", "issue:42490|issue:43696", + "issue:29159|issue:32090", "issue:28218|issue:34567", "issue:44977|issue:46701", "issue:45084|issue:45520", @@ -8771,198 +8764,209 @@ "issue:43295|issue:43873", "issue:44186|issue:46192", "issue:45823|issue:46192", - "issue:43295|issue:44610" + "issue:43295|issue:44610", + "issue:43572|issue:46821", + "issue:43550|issue:43844", + "issue:45200|issue:45823", + "issue:29870|issue:29942", + "issue:45520|issue:46178" ], - "prepared_review_unit_hash": "904fc3f5e9e6990ccae5046070c8436a2d9776e1fc13e818951ccaccd2ea80e3", + "prepared_review_unit_hash": "92e174fde8eba83be6fc6d04e23c713543aaea116878e240f3d40d82434d4461", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35864, - "estimated_input_tokens": 8966, - "estimated_eval_tokens": 18188 + "soft_pair_count": 28, + "serialized_chars": 36238, + "estimated_input_tokens": 9060, + "estimated_eval_tokens": 18376 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: most pairs share only broad subsystem vocabulary (tokenizers, processors, attention, serve, multimodal models) but do not describe the same concrete bug. I would triage it around the processor/tokenizer regression issue, but I would not merge any of the suggested soft pairs.", - "confidence": 0.89, - "canonical_issue_reason": "issue:43295 is the best hub in the soft-similarity graph and the most representative user-facing regression around processor/tokenizer behavior. It also has the strongest overlap with other processor-related reports in this set, even though the candidates are not true duplicates.", + "summary": "The soft pairs are mostly false positives: they share broad subsystems like tokenizers, serving, or multimodal models, but not the same concrete bug. No PRs are present.", + "confidence": 0.86, + "canonical_issue_reason": "issue:29159 is the broadest, most reusable tokenizer-behavior report among the candidates and the best umbrella-style issue, even though it does not match the soft pairs as duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43295 is the strongest single issue to represent the cluster because it is recent, concrete, and sits at the center of the processor/tokenizer-related soft links. It is the closest thing to a unifying bug among otherwise unrelated reports.", + "best_issue_reason": "issue:29159 is the most general and actionable issue to anchor this set; it is well-scoped and likely the most useful canonical representative if these were being grouped loosely by tokenizer behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45229", - "right": "issue:45823", - "accept": false, - "reason": "Same model family and inference context, but one is CUDA OOM and the other is a device-mapping mismatch; different failures and different fixes." - }, - { - "left": "issue:44869", - "right": "issue:46752", - "accept": false, - "reason": "Whisper timestamp decoding crash vs empty-chat-template IndexError are unrelated code paths." - }, - { - "left": "issue:43994", - "right": "issue:44493", - "accept": false, - "reason": "One is a SigLIP2 quality/regression report and the other is a position-id key warning affecting many models; not the same bug." - }, - { - "left": "issue:45588", - "right": "issue:46050", - "accept": false, - "reason": "Flash-attention sink-less-model crash vs sinq save/load breakage are different issues in different code paths." - }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 test failures and MLflow job state reporting are unrelated." + "reason": "Different bugs: Flash Attention 2 test failures vs MLflow reporting after training failure." }, { - "left": "issue:29159", - "right": "issue:32090", + "left": "issue:42490", + "right": "issue:43696", "accept": false, - "reason": "Tokenizer fast/slow inconsistency is unrelated to a NoneType broadcast error in Trainer." + "reason": "Unrelated: local-path import warning vs CUDA OOM on GPT-oss-20b." }, { - "left": "issue:42490", - "right": "issue:43696", + "left": "issue:29159", + "right": "issue:32090", "accept": false, - "reason": "Path-trailing-slash import warning and GPT-oss CUDA OOM are unrelated." + "reason": "Tokenizer slow/fast inconsistency is unrelated to a Trainer NoneType GPU broadcast error." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Added-token spacing bug and TrainerState token counter bug are different tokenizer/training-state problems." + "reason": "Tokenizer added-space behavior and TrainerState token counting are different code paths and symptoms." }, { "left": "issue:44977", "right": "issue:46701", "accept": false, - "reason": "Qwen flash-attention generation failure and DiffusionGemma pipeline failure are different model/code paths." + "reason": "Both involve generation, but different model families and failure modes; not the same bug." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Template compilation error and Python 3.13 flash_attn import key error are unrelated." + "reason": "Non-template-node compile error is unrelated to flash_attn import-key errors." }, { "left": "issue:43065", "right": "issue:46369", "accept": false, - "reason": "Both mention SAM3, but one is about a dummy Conv2d design choice and the other is a memory overflow report." + "reason": "Same model family, but dummy Conv2d construction and memory overflow are not the same defect." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness and flex_attention import version checking are different bugs." + "reason": "Attention correctness on MPS is unrelated to a PyTorch version check bug in flex_attention imports." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp export failure and SigLIP2 zero probabilities are unrelated model issues." + "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero probabilities." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Both are serve/runtime bugs, but one is a multimodal continuous-batching crash and the other is KV-cache continuation indexing; not the same failure." + "reason": "Both are serve crashes, but they hit different paths: multimodal tensor conversion vs KV-cache continuation indexing." }, { "left": "issue:41720", "right": "issue:46710", "accept": false, - "reason": "CUDA assert on A800 with auto device mapping vs accuracy regression after upgrade are unrelated." + "reason": "Different models and problems: cudaErrorAssert with auto device mapping vs an accuracy regression after upgrade." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are distinct issues." + "reason": "SigLIP2 parity discrepancy and resume_from_checkpoint key mismatch are unrelated." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4Processor _tokenizer crash in serve and rope-parameter list type error are unrelated." + "reason": "Gemma4Processor attribute crash and rope-param JSON list handling are different bugs." }, { "left": "issue:43404", "right": "issue:46710", "accept": false, - "reason": "Mistral3 lm_head tying bug and DeepSeek accuracy regression are different model correctness issues." + "reason": "Weight tying in Mistral3 is unrelated to DeepSeek accuracy regression." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "Bfloat16 dtype mismatch in inference and Wav2Vec2 save/tokenization failure are unrelated." + "reason": "bfloat16 dtype mismatch tests are unrelated to save_pretrained/tokenization failures." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Both involve processors, but one is a tokenizer/processor API regression and the other is a batched video-shape bug." + "reason": "Processor-tokenizer regression and batched video shape bug are different issues." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "Flash-attention sink-less crash and multimodal config unwrapping are different code paths." + "reason": "Flash-attention sink handling and Qwen multimodal config unwrapping are separate code paths." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Flash-attention import symbol error and meta-device loading breakage are not the same bug." + "reason": "Import name change in flash_attention_utils is unrelated to meta-device model loading failures." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Video-input ValueError in Qwen3OmniMoe and Gemma4 text-only mm_token_type_ids defaulting are unrelated multimodal issues." + "reason": "Video processing ValueError and Gemma4 mm_token_type_ids defaulting are unrelated." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Processor/tokenizer regression and quantization offloading behavior are not the same defect." + "reason": "Processor API regression is unrelated to quantization offloading behavior." }, { "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "LayoutLMv2 NER/padding crash and return_tensors being ignored are different tokenizer API bugs." + "reason": "Both mention tokenizers, but NER/padding crashes and ignored return_tensors are distinct bugs." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Device-map forward mismatch and return_tensors handling are unrelated." + "reason": "Gemma4 device mismatch is unrelated to return_tensors being ignored." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Processor/tokenizer regression and OmDet-Turbo input-size mismatch are both processor-related but describe different concrete failures." + "reason": "Processor/tokenizer regression and OmDet-Turbo image size mismatch are not the same bug." + }, + { + "left": "issue:43572", + "right": "issue:46821", + "accept": false, + "reason": "Different config/attribute issues in unrelated models." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile SDPA failure and ZeRO-3 gradient growth are unrelated training/runtime bugs." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but text-only fine-tuning defaults and forward device mismatch are different failures." + }, + { + "left": "issue:29870", + "right": "issue:29942", + "accept": false, + "reason": "Bart generate masking error is unrelated to Flash Attention 2 test failures." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "flash_attn import error on Python 3.13 and BACKENDS_MAPPING tensorflow_text error are different backend-registration issues." } ] }, @@ -8987,8 +8991,6 @@ "cluster_id": "cluster-10000-502", "nodes": [ "issue:28218", - "issue:29870", - "issue:29942", "issue:35141", "issue:37428", "issue:39692", @@ -8998,15 +9000,16 @@ "issue:43262", "issue:43278", "issue:43295", + "issue:43334", "issue:43450", "issue:43452", "issue:43493", - "issue:43550", + "issue:43525", "issue:43572", "issue:43575", "issue:43653", + "issue:43819", "issue:43824", - "issue:43844", "issue:43856", "issue:43950", "issue:43975", @@ -9016,7 +9019,9 @@ "issue:44521", "issue:44554", "issue:44737", + "issue:44792", "issue:44849", + "issue:44898", "issue:44987", "issue:45127", "issue:45198", @@ -9024,23 +9029,17 @@ "issue:45290", "issue:45356", "issue:45381", - "issue:45520", - "issue:45823", - "issue:46178", + "issue:45412", + "issue:45563", "issue:46192", "issue:46209", "issue:46285", "issue:46512", - "issue:46821", "issue:46894", - "issue:46899" + "issue:46899", + "issue:46932" ], "soft_pairs": [ - "issue:43572|issue:46821", - "issue:43550|issue:43844", - "issue:45200|issue:45823", - "issue:29870|issue:29942", - "issue:45520|issue:46178", "issue:43450|issue:43575", "issue:43572|issue:44987", "issue:44554|issue:45356", @@ -9056,211 +9055,219 @@ "issue:43278|issue:43856", "issue:28218|issue:35141", "issue:44737|issue:46894", - "issue:43950|issue:46285", "issue:43824|issue:43975", + "issue:43950|issue:46285", "issue:42915|issue:43824", "issue:43452|issue:46512", "issue:43262|issue:43450", "issue:44442|issue:45198", - "issue:45200|issue:46899" + "issue:45200|issue:46932", + "issue:45200|issue:46899", + "issue:43334|issue:43525", + "issue:44792|issue:45563", + "issue:44898|issue:45412", + "issue:43819|issue:44112" ], - "prepared_review_unit_hash": "336f8bdcff7d10ddcdc7b00281bd4917272c0477dbb2865928c925be5b3355e4", + "prepared_review_unit_hash": "f28cf8134459d443aae546a1083c702edc67f1492e761e368abf5ce4b86a3645", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36170, - "estimated_input_tokens": 9043, - "estimated_eval_tokens": 18342 + "serialized_chars": 36188, + "estimated_input_tokens": 9047, + "estimated_eval_tokens": 18350 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is highly heterogeneous: tokenizer quirks, generation bugs, multimodal processor regressions, import errors, and training/runtime issues. I don\u2019t see any soft pair that clearly describes the same underlying bug or change, so I reject all soft edges.", - "confidence": 0.42, - "canonical_issue_reason": "There is no single obvious duplicate family spanning the cluster. If a representative anchor is needed, issue 43950 is the broadest framework-level regression among the candidates (a core from_pretrained corruption bug).", + "summary": "The set is mostly heterogeneous; almost all soft edges are false positives. The only plausible duplicate pair is the two missing-`pad_token_id` config-loading regressions.", + "confidence": 0.84, + "canonical_issue_reason": "If a canonical issue must be chosen, issue:43334 is the best anchor because it captures the shared `pad_token_id`-missing regression that issue:43525 appears to duplicate; the rest of the items are unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43950 is the best standalone canonical candidate because it is a core loading regression with silent corruption, making it broader and more actionable than the more model-specific or documentation-oriented issues here.", + "best_issue_reason": "issue:43334 is the strongest representative of the only real duplicate subcluster (shared `pad_token_id` AttributeError during model loading).", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43572", - "right": "issue:46821", - "accept": false, - "reason": "Both are configuration/attribute problems, but they affect different components and different missing fields; not the same bug." - }, - { - "left": "issue:43550", - "right": "issue:43844", - "accept": false, - "reason": "One is a torch.compile/SDPA model failure and the other is a ZeRO-3 gradient issue; different code paths and symptoms." - }, - { - "left": "issue:45200", - "right": "issue:45823", - "accept": false, - "reason": "Gemma4 text-only mm_token_type_ids defaulting and device_map PLE mismatch are distinct forward-path bugs." - }, - { - "left": "issue:29870", - "right": "issue:29942", - "accept": false, - "reason": "Bart generate decoder_attention_mask initialization and Flash Attention 2 test failures are unrelated." - }, - { - "left": "issue:45520", - "right": "issue:46178", - "accept": false, - "reason": "These are separate import/backend lookup failures involving different backends and modules." - }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video batch shape handling and Qwen2-57B tensor-parallel OOM are different problems." + "reason": "Different problems: batched video shape vs tensor-parallel OOM." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "A missing config attribute versus a model-loading failure for a different repo; not the same underlying issue." + "reason": "Both are loading failures, but for different models and different root causes." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "An MPS attention correctness bug and a tokenizer codec regression are unrelated." + "reason": "MPS attention correctness bug vs tokenizer regression; unrelated code paths." }, { "left": "issue:42490", "right": "issue:43493", "accept": false, - "reason": "Local-path import warning and SigLIP2 implementation discrepancy do not share the same code path." + "reason": "Local-path slash warning is unrelated to SigLIP2 implementation discrepancy." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Different ImportErrors from different symbols; no evidence of a shared root cause." + "reason": "Different ImportErrors involving different symbols and code paths." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "Documentation example issues are not the same as silently ignored return_tensors behavior." + "reason": "Documentation-example issues vs `return_tensors` handling; not the same bug." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to Qwen3.5 hidden-states behavior." + "reason": "Generation kwargs update bug vs Qwen3.5 hidden-states bug; different failures." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "Both touch tied weights, but one is serialization inconsistency and the other is LoRA merge collapse; different concrete bugs." + "reason": "Both involve tied weights, but one is serialization and the other is LoRA merge collapse." }, { "left": "issue:43295", "right": "issue:43493", "accept": false, - "reason": "A processor/tokenizer API regression is not the same as the SigLIP2 HF-vs-JAX discrepancy." + "reason": "Processor/tokenizer regression vs SigLIP2 implementation mismatch." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both are apply_chat_template-related, but one is assistant mask generation for multimodal inputs and the other is a crash on tool-call messages." + "reason": "Both touch `apply_chat_template`, but one is zeroed assistant masks and the other is a crash on tool-call messages." }, { "left": "issue:45381", "right": "issue:46209", "accept": false, - "reason": "Qwen2.5-VL vision_position_ids and FP8 B200 inference are unrelated issues." + "reason": "Video `vision_position_ids` bug vs FP8 inference failure; unrelated." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "Tokenizer special-token registration and a GraniteSpeech CI test failure are not the same bug." + "reason": "Tokenizer special-token registration issue vs GraniteSpeech CI device override test." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "A dtype change between train/eval and MoE training memory inefficiency are different problems." + "reason": "dtype drift in evaluation vs MoE training memory usage." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer spacing after added tokens and embedding reinitialization after resize are distinct behaviors." + "reason": "Tokenizer spacing bug vs embedding reinitialization during resize; different bugs." }, { "left": "issue:44737", "right": "issue:46894", "accept": false, - "reason": "CPU placement in XLNet positional encoding and CohereASR double-shift training loss are unrelated." + "reason": "XLNet CPU tensor allocation issue vs ASR double-shift training-loss bug." }, { - "left": "issue:43950", - "right": "issue:46285", + "left": "issue:43824", + "right": "issue:43975", "accept": false, - "reason": "Non-persistent buffer corruption in from_pretrained and CLIP text model structure inconsistency are different regressions." + "reason": "Missing export ImportError vs incorrect detokenization; unrelated." }, { - "left": "issue:43824", - "right": "issue:43975", + "left": "issue:43950", + "right": "issue:46285", "accept": false, - "reason": "An import failure and a detokenization bug are not the same underlying issue." + "reason": "Non-persistent buffer corruption vs CLIP text-model structure inconsistency." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Qwen3Moe FP8 failure and a Qwen2.5-VL import error share only a broad model-family context, not a bug." + "reason": "FP8-config model failure vs missing Qwen2.5-VL import; different code paths." }, { "left": "issue:43452", "right": "issue:46512", "accept": false, - "reason": "GGUF loading breaks and stale config.dtype after from_config are different lifecycle bugs." + "reason": "`gguf_file` loading breakage vs stale `config.dtype` after `from_config`." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio chat-template sampling-rate defaults and batched video shape handling are unrelated processor bugs." + "reason": "Audio chat-template sample-rate default vs batched video output shape; different processors." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Different speech-model failures with different failure modes; not one concrete bug." + "reason": "Tokenizer auto-loading failure vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:45200", + "right": "issue:46932", + "accept": false, + "reason": "Missing multimodal token-type defaults vs quantized embedder dtype casting; not the same bug." }, { "left": "issue:45200", "right": "issue:46899", "accept": false, - "reason": "Gemma4 mm_token_type_ids defaults and unsafe dtype casts in multimodal embedders are different regressions." + "reason": "Text-only `mm_token_type_ids` defaulting vs unsafe dtype cast in quantized embedders." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": true, + "reason": "Both report the same `pad_token_id`-missing AttributeError in model/config loading, likely fixable with one shared fallback or config change." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure vs stale paged-generate warning." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver interpolation failure vs RT-DETR memory release issue." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC forward/from_latents mismatch vs GraniteSpeech CI device override regression." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is conservative and grounded in the packet. The false soft-edge verdicts mostly stay at the title/symptom level and do not overclaim shared root causes. The only accepted pair, issue:43334 with issue:43525, is supported by the shared `pad_token_id` AttributeError symptom in the titles, and the rest of the set does look heterogeneous. No mergeability concern here since these are issue pairs, not PR pairs." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -9284,11 +9291,10 @@ "issue:37428", "issue:38175", "issue:41720", - "issue:43334", "issue:43335", "issue:43388", "issue:43425", - "issue:43525", + "issue:43504", "issue:43526", "issue:43550", "issue:43606", @@ -9299,22 +9305,24 @@ "issue:43901", "issue:44112", "issue:44206", + "issue:44448", "issue:44466", "issue:44479", "issue:44561", "issue:44610", - "issue:44792", + "issue:44743", + "issue:44855", "issue:44857", - "issue:44898", "issue:44995", "issue:45083", "issue:45200", "issue:45325", "issue:45412", - "issue:45563", + "issue:45520", "issue:45584", "issue:45593", "issue:45759", + "issue:46170", "issue:46192", "issue:46209", "issue:46234", @@ -9322,15 +9330,9 @@ "issue:46424", "issue:46512", "issue:46821", - "issue:46858", - "issue:46932" + "issue:46858" ], "soft_pairs": [ - "issue:45200|issue:46932", - "issue:44792|issue:45563", - "issue:43334|issue:43525", - "issue:44898|issue:45412", - "issue:43819|issue:44112", "issue:44479|issue:46192", "issue:45200|issue:45325", "issue:41720|issue:46209", @@ -9353,204 +9355,202 @@ "issue:43550|issue:44112", "issue:28282|issue:33357", "issue:37428|issue:43425", - "issue:43611|issue:44466" + "issue:43611|issue:44466", + "issue:44448|issue:44743", + "issue:43335|issue:43504", + "issue:45520|issue:46170", + "issue:44855|issue:45083" ], - "prepared_review_unit_hash": "8b0e2f902605f3e38920579a5d6058a5372fd7f634a2e71a6c1bcec291105851", + "prepared_review_unit_hash": "67e3afaa342c1607671de2937cbc2f09d9586924fcc16107824af67e34353c05", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35854, - "estimated_input_tokens": 8964, - "estimated_eval_tokens": 18184 + "soft_pair_count": 27, + "serialized_chars": 35640, + "estimated_input_tokens": 8910, + "estimated_eval_tokens": 18076 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose bag of unrelated transformer bugs, regressions, and docs/CI issues. None of the soft-linked pairs appear to be the same underlying defect or change, so there is no good canonical duplicate target.", - "confidence": 0.97, + "summary": "These are mostly unrelated bug reports across different models, APIs, and environments. I did not find any soft pair that clearly describes the same underlying bug/change, so no duplicates were accepted and no canonical artifact stands out.", + "confidence": 0.96, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45200", - "right": "issue:46932", - "accept": false, - "reason": "Both are Gemma4-related, but they describe different failures: missing default mm_token_type_ids vs dtype casting for quantized embedders." - }, - { - "left": "issue:44792", - "right": "issue:45563", - "accept": false, - "reason": "Janus image-generation test failure and paged-generate stale warning are separate generate-path issues." - }, - { - "left": "issue:43334", - "right": "issue:43525", - "accept": false, - "reason": "Both mention missing pad_token_id, but on different config classes and model families; not the same bug." - }, - { - "left": "issue:44898", - "right": "issue:45412", - "accept": false, - "reason": "Perceiver interpolation failure and RT-DETR memory release are unrelated model/runtime problems." - }, - { - "left": "issue:43819", - "right": "issue:44112", - "accept": false, - "reason": "DAC forward/from_latents mismatch is a model implementation bug, while GraniteSpeech is a stale CI device-override test." - }, { "left": "issue:44479", "right": "issue:46192", "accept": false, - "reason": "Video input regression in Qwen multimodal models and return_tensors being ignored are different API/code paths." + "reason": "Both concern multimodal input handling, but one is a video regression in several Qwen VL models while the other is a return_tensors/text_kwargs precedence bug; different code paths." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma4 mm_token_type_ids handling and Qwen2.5-VL rope-index scaling are separate multimodal bugs." + "reason": "One is about missing mm_token_type_ids defaults for text-only fine-tuning; the other is a rope/position_ids scaling bug in Qwen2.5-VL. Different behavior and fix surface." }, { "left": "issue:41720", "right": "issue:46209", "accept": false, - "reason": "Different hardware/backend failures: cudaErrorAssert on A800 vs FP8 inference failure on B200." + "reason": "Both are device/hardware failures, but one is a cudaErrorAssert on A800 with Qwen3 auto device mapping and the other is FP8 inference failure on B200. Not the same bug." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs issues, but they concern different pipeline docs and different deprecated behaviors." + "reason": "Both are docs issues, but they cover different APIs and deprecations: pipeline() removal versus return_all_scores wording." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug and Bamba torch.compile/SDPA failure are unrelated." + "reason": "Different model families and failure modes: SwitchTransformers config creates an unwanted sparse layer, while Bamba fails under torch.compile with SDPA." }, { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "Both involve missing utilities, but they are different symbols with different breakages and affected models." + "reason": "Both mention missing import_utils helpers, but they are different symbols affecting different models and dependencies." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC latent mismatch and qwen3_omni_moe feature-length helper behavior are unrelated." + "reason": "DAC.from_latents vs qwen3_omni_moe feature-length helper are unrelated model internals and bugs." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Tuple-label dropping in gather_for_metrics and BeitImageProcessorFast reduce_labels are different label-processing bugs." + "reason": "Both involve labels, but one is gather_for_metrics truncating the last batch and the other is BeitImageProcessorFast.reduce_labels returning only one label. Different root causes." }, { "left": "issue:45759", "right": "issue:46285", "accept": false, - "reason": "Composite Qwen config unwrapping and CLIP text-model structure inconsistency are different architectural issues." + "reason": "One is an AutoModelForCausalLM config-unwrapping issue for Qwen multimodal configs; the other is a CLIP text model structure inconsistency." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "AMP crash in LwDetrImageLoss and D-FINE auxiliary-loss behavior are unrelated." + "reason": "Different detection/training bugs: float16 AMP crash in LwDetrImageLoss versus missing auxiliary losses in D-FINE when denoising is off." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Processor input-size mismatch and float16/CUDA loss crash are different bugs." + "reason": "Different models and symptoms: OmDet-Turbo processor output size mismatch versus LwDetrImageLoss AMP crash." }, { "left": "issue:46424", "right": "issue:46858", "accept": false, - "reason": "Max-cache-length handling and static-cache seq-length tensor arithmetic are distinct cache issues." + "reason": "Both are cache-related, but one is max_cache_length only applying to Quantized Cache while the other is get_seq_length returning a device tensor that breaks slice arithmetic under torch.compile." }, { "left": "issue:44995", "right": "issue:46858", "accept": false, - "reason": "Stale indexer cache on second forward is unrelated to device-tensor seq-length arithmetic in static cache." + "reason": "Both mention cache behavior, but one is a stale indexer cache causing second-forward crashes and the other is a device-tensor seq_length arithmetic bug." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "CPU offload device mismatch and qwen3_omni_moe length helper behavior are unrelated." + "reason": "CPU offload device mismatch in bark-small is unrelated to qwen3_omni_moe feature-extraction length behavior." }, { "left": "issue:46512", "right": "issue:46821", "accept": false, - "reason": "Stale config dtype after from_config and TimesFM window_size AttributeError are different configuration bugs." + "reason": "Different API regressions: stale config.dtype after from_config(dtype=...) versus AttributeError for TimesFM 2.5 window_size." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "Unsupported center arg in feature extraction and Whisper empty-transcription generation failure are unrelated." + "reason": "LasrFeatureExtractor center-arg crash and Whisper empty-transcription failure are distinct preprocessing/generation bugs." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "Zero probabilities in SigLIP2 and RT-DETR memory cleanup are different model/runtime issues." + "reason": "Unrelated symptoms and components: zero probabilities in SigLIP2 versus RT-DETR memory not being released." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and BigBirdTokenizer special-token registration are unrelated." + "reason": "Different areas: SwitchTransformers sparse-layer creation versus BigBirdTokenizer special-token registration and decode output." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are separate problems." + "reason": "Flash Attention 2 test failures and _gpu_broadcast_one NoneType errors are unrelated test/runtime issues." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "Bamba torch.compile SDPA failure and GraniteSpeech stale override test are not the same code-path bug." + "reason": "Bamba SDPA/torch.compile failure and GraniteSpeech stale device-override CI failure do not share the same code path." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "Missing PyTorch ImportError and MacOS bus error on a CLIP model are unrelated environment/model failures." + "reason": "ImportError about missing PyTorch and a MacOS bus error in CLIP model loading are different failure modes." }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "Missing flash-attention helper import and Torch 2.10 incompatibility are different versioning issues." + "reason": "Importing a missing flash-attention helper and a Torch 2.10 compatibility report are too broad and not the same bug." }, { "left": "issue:43611", "right": "issue:44466", "accept": false, - "reason": "base_model_prefix loading break and lm_head.weight serialization inconsistency are separate model-loading/serialization bugs." + "reason": "base_model_prefix loading breakage and lm_head.weight serialization inconsistency are separate v5 regressions." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output differences and recurrent state reset in modular_qwen3_5 are unrelated model behavior changes." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Different bugs in different models: sparse layer creation in SwitchTransformers versus legacy field loading in BeitForSemanticSegmentation." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Python 3.13 flash_attn import-key issue and Windows ROCm torch.distributed import/type errors are different platform-specific problems." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "DebertaV2Model import parsing on Python 3.13 is unrelated to qwen3_omni_moe feature-length helper behavior." } ] }, @@ -9579,6 +9579,7 @@ "issue:29127", "issue:31501", "issue:32090", + "issue:33357", "issue:34689", "issue:35141", "issue:37428", @@ -9586,8 +9587,8 @@ "issue:42490", "issue:43278", "issue:43335", - "issue:43504", "issue:43526", + "issue:43575", "issue:43606", "issue:43701", "issue:43720", @@ -9595,16 +9596,16 @@ "issue:43825", "issue:43854", "issue:44112", + "issue:44206", "issue:44351", "issue:44448", "issue:44493", "issue:44514", "issue:44521", "issue:44617", - "issue:44743", "issue:44779", - "issue:44855", "issue:44857", + "issue:45072", "issue:45083", "issue:45127", "issue:45245", @@ -9612,11 +9613,10 @@ "issue:45310", "issue:45459", "issue:45478", - "issue:45520", "issue:45588", "issue:45593", + "issue:45823", "issue:45859", - "issue:46170", "issue:46234", "issue:46326", "issue:46431", @@ -9624,26 +9624,22 @@ "issue:46697" ], "soft_pairs": [ - "issue:44448|issue:44743", - "issue:43335|issue:43504", - "issue:45520|issue:46170", - "issue:44855|issue:45083", "issue:43278|issue:43825", "issue:43335|issue:45083", "issue:43335|issue:45593", - "issue:29127|issue:32090", "issue:43335|issue:44112", - "issue:44514|issue:44521", + "issue:29127|issue:32090", "issue:43720|issue:44112", + "issue:44514|issue:44521", "issue:45290|issue:46326", "issue:45588|issue:45859", "issue:42490|issue:43854", "issue:37428|issue:44351", "issue:34689|issue:35141", "issue:45127|issue:45245", + "issue:39692|issue:43526", "issue:43819|issue:44493", "issue:46431|issue:46566", - "issue:39692|issue:43526", "issue:26413|issue:31501", "issue:44448|issue:44779", "issue:45310|issue:45478", @@ -9652,216 +9648,220 @@ "issue:45459|issue:46234", "issue:39692|issue:44493", "issue:43701|issue:45083", - "issue:44617|issue:46697" + "issue:44617|issue:46697", + "issue:44206|issue:45823", + "issue:32090|issue:33357", + "issue:42490|issue:43575", + "issue:44857|issue:45072" ], - "prepared_review_unit_hash": "f4771eee44c98543189d8be62cd516224a6bf6cd417c4bf53451c0a759bd188d", + "prepared_review_unit_hash": "7d0a761909637c5fe47feea7d6f433b1a00896d3a3888c0f34d3a9dcfa7ec8af", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36644, - "estimated_input_tokens": 9161, - "estimated_eval_tokens": 18578 + "serialized_chars": 36496, + "estimated_input_tokens": 9124, + "estimated_eval_tokens": 18504 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous with several unrelated issue pairs. The only clear duplicate pair is the two identical Qwen3.5 MoE from_pretrained issues; the rest should stay separate.", - "confidence": 0.86, - "canonical_issue_reason": "Issue 45310 is the cleanest canonical issue because it is the earlier of the exact duplicate pair and captures the same Qwen3.5 MoE from_pretrained failure as 45478.", + "summary": "This cluster is mostly a set of unrelated bug reports; only one clear duplicate pair stands out. The strongest match is the duplicated Qwen3.5 MoE from_pretrained regression (45310/45478).", + "confidence": 0.93, + "canonical_issue_reason": "issue:45310 is the best canonical issue because it has an exact duplicate partner (issue:45478) with the same title and the same underlying Qwen3.5 MoE from_pretrained failure, and it appears earlier in time.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45310 is the best global issue anchor: it is a true duplicate target, has a precise title, and is the only pair in this set that clearly represents the same bug.", + "best_issue_reason": "issue:45310 is the most representative issue in the cluster: it captures the clearest duplicate theme with the strongest textual match and enough discussion to serve as the anchor report.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44448", - "right": "issue:44743", + "left": "issue:43278", + "right": "issue:43825", "accept": false, - "reason": "Different problems: Pegasus output parity vs Qwen3 recurrent-state reset in a modular file; no shared concrete bug path." + "reason": "Different problems: one is a dtype change during train/eval, the other is a pipeline error message about translation support." }, { "left": "issue:43335", - "right": "issue:43504", + "right": "issue:45083", "accept": false, - "reason": "Different models and failures: SwitchTransformers config construction vs Beit pretrained loading with a legacy field." + "reason": "Different code paths: SwitchTransformers sparse-layer construction vs qwen3_omni_moe feature-length helper behavior." }, { - "left": "issue:45520", - "right": "issue:46170", + "left": "issue:43335", + "right": "issue:45593", "accept": false, - "reason": "Both are import-related, but the causes and platforms differ: flash_attn KeyError on Python 3.13 vs Windows/ROCm torch.distributed dependency issues." + "reason": "Unrelated issues: sparse layer creation in SwitchTransformers vs missing auxiliary losses in D-FINE." }, { - "left": "issue:44855", - "right": "issue:45083", + "left": "issue:43335", + "right": "issue:44112", "accept": false, - "reason": "Unrelated bugs: Python 3.13 IndentationError in DebertaV2 import vs qwen3_omni_moe helper length behavior." - }, - { - "left": "issue:43278", - "right": "issue:43825", - "accept": false, - "reason": "Different symptoms and subsystems: embedding dtype drift during eval vs pipeline translation warning messaging." - }, - { - "left": "issue:43335", - "right": "issue:45083", - "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to qwen3_omni_moe feature-length helper behavior." - }, - { - "left": "issue:43335", - "right": "issue:45593", - "accept": false, - "reason": "Different models and code paths: SwitchTransformers config vs D-FINE auxiliary-loss handling." + "reason": "No concrete overlap: SwitchTransformers config bug vs a GraniteSpeech CI device override test failure." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "Different error domains: LayoutLMv3 box validation messaging vs Trainer/_gpu_broadcast_one NoneType handling." + "reason": "Different symptoms and subsystems: LayoutLMv3 error messaging vs a Trainer/_gpu_broadcast_one NoneType type error." }, { - "left": "issue:43335", + "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "No shared bug: SwitchTransformers sparse layer creation vs GraniteSpeech stale device override CI failure." + "reason": "Different bugs in different components; no shared failing code path." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both involve chat templates, but one is a batched padding=False crash and the other is incorrect assistant_masks for multimodal inputs; not the same concrete bug." - }, - { - "left": "issue:43720", - "right": "issue:44112", - "accept": false, - "reason": "Different CI/runtime failures: unpacking packed BitNet weights during accelerate loading vs GraniteSpeech device override test staleness." + "reason": "Both involve apply_chat_template, but one is a batched padding=False crash and the other is incorrect all-zero assistant masks for multimodal inputs; not the same concrete bug." }, { "left": "issue:45290", "right": "issue:46326", "accept": false, - "reason": "Both touch chat templating, but one crashes on tool-call assistant messages with no content while the other is a mistral_common continue_final_message tokenizer bug." + "reason": "Both touch chat templating, but one is tool-call handling with empty assistant content and the other is continue_final_message behavior for mistral_common tokenizers." }, { "left": "issue:45588", "right": "issue:45859", "accept": false, - "reason": "Different failures in different places: flash_attention integration s_aux=None crash vs Qwen3_5MoeTextRotaryEmbedding CPU offload incompatibility." + "reason": "Different failure modes and code paths: flash_attention sink-less s_aux handling vs Qwen3_5MoeTextRotaryEmbedding CPU-offload compatibility." }, { "left": "issue:42490", "right": "issue:43854", "accept": false, - "reason": "Both are loading-related, but one is a path-with-trailing-slash import warning and the other is a GLM-4.7-Flash unit-test loading failure." + "reason": "Not the same bug: local-path-with-slash import warning vs a unit-test model loading failure for GLM-4.7-Flash." }, { "left": "issue:37428", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and modules: _flash_supports_window_size vs HybridCache." + "reason": "Both are import errors, but they concern different missing symbols in different modules." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Different model-loading issues: Llama 3.2 Vision breakage vs embedding reinitialization after resize when tie_word_embedding is False." + "reason": "Different regressions: Llama 3.2 Vision model loading vs post_init reinitializing output embeddings after resize when tie_word_embedding is False." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "Unrelated: LoRA merge collapse with extended vocab vs category-count limit RuntimeError." + "reason": "Completely different bugs: LoRA/extended-vocab collapse vs category-count overflow." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "Different model areas and failures: SigLIP2 docs/model-processor issues vs BEiT image processor label reduction bug." }, { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "Different bugs: DAC from_latents mismatch due to missing STE vs unexpected key with position id across many models." + "reason": "Different concrete problems: DAC from_latents mismatch vs unexpected position-id key warnings across many models." }, { "left": "issue:46431", "right": "issue:46566", "accept": false, - "reason": "Different subsystems and symptoms: missing chat template on a processor vs DiffusionGemma weights staying on meta with device_map=auto." - }, - { - "left": "issue:39692", - "right": "issue:43526", - "accept": false, - "reason": "Docs/example errors for SigLIP2 are unrelated to BeitImageProcessorFast reduce_labels returning only one label." + "reason": "No shared code path: missing chat template on a processor vs DiffusionGemma staying on meta with device_map='auto'." }, { "left": "issue:26413", "right": "issue:31501", "accept": false, - "reason": "Resume-from-checkpoint epoch_iterator failure is unrelated to FineWeb SLM training not starting." + "reason": "One is a resume_from_checkpoint epoch-iterator failure; the other is a FineWeb training startup issue with different root causes." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Both mention output discrepancies, but they affect different models/tokenizers and are not the same underlying bug." + "reason": "Both are v5 regressions, but one is Pegasus output drift and the other is Deepseek tokenizer correctness; different artifacts and fixes." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Exact duplicate: same title, same Qwen3.5 MoE from_pretrained error on transformers>=5.4.0." + "reason": "Exact duplicate reports of the same Qwen3.5 MoE from_pretrained error, with matching title and underlying failure." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "Different failures: CPU offload device mismatch in bark-small CI vs LwDetrImageLoss float16 AMP CUDA crash." + "reason": "Different models and failures: CPU offload device mismatch in bark-small vs AMP CUDA crash in LwDetrImageLoss." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer added-space behavior is unrelated to Trainer NoneType broadcast errors." + "reason": "Tokenizer spacing bug vs Trainer NoneType broadcast error; unrelated." }, { "left": "issue:45459", "right": "issue:46234", "accept": false, - "reason": "Different import-utils issues: protobuf decode-error masking tokenizer errors vs missing is_falcon_mamba_ssm_available helper." + "reason": "Different issues: tokenizer error masking when protobuf is missing vs a missing availability helper in import_utils." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 example/documentation failures are unrelated to widespread position-id key warnings." + "reason": "Different subjects: SigLIP2 documentation/model setup issues vs unexpected position-id keys across many models." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Resume_from_checkpoint key mismatch is unrelated to qwen3_omni_moe feature-length helper behavior." + "reason": "Both are bug reports, but one is resume_from_checkpoint key mismatch and the other is a qwen3_omni_moe feature-length helper issue." }, { "left": "issue:44617", "right": "issue:46697", "accept": false, - "reason": "Different image/video model issues: Sam3Video CUDA OOM vs Swin2SR over-padding when size_divisor already matches." + "reason": "Different performance/correctness problems: CUDA OOM in Sam3Video vs over-padding in Swin2SR image preprocessing." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Different issues in different models/components: unsupported center arg in LasrFeatureExtractor vs Gemma4 PLE device mismatch." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Different failure modes and platforms: NoneType broadcast type error vs MacOS bus error with a CLIP model." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Different bugs: path-ending-with-slash import warning vs tensor-parallel load OOM." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Both are dtype-related, but they affect different models and code paths; not the same concrete bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded and conservative overall. The only accepted duplicate pair is clearly supported by the packet: issue:45310 and issue:45478 have identical titles and matching Qwen3.5 MoE from_pretrained failure descriptions. The other soft-edge rejections are stated cautiously and do not overclaim shared root causes." + "feedback": "The summary is well grounded in the packet. The only accepted duplicate pair, issue:45310/issue:45478, is strongly supported by the identical title and high textual similarity, and the other soft-edge rejections are phrased conservatively without overclaiming cross-issue sameness. The canonical choice is also reasonable given the earlier timestamp." }, "evaluator_used": true, "retried": false, @@ -9884,12 +9884,12 @@ "nodes": [ "issue:10000", "issue:15354", + "issue:25251", "issue:27957", "issue:29870", "issue:29942", "issue:32090", "issue:33290", - "issue:33357", "issue:34689", "issue:35141", "issue:37428", @@ -9902,28 +9902,28 @@ "issue:43381", "issue:43526", "issue:43550", - "issue:43575", "issue:43873", "issue:43927", "issue:43976", - "issue:44206", "issue:44360", "issue:44448", + "issue:44466", "issue:44493", "issue:44561", + "issue:44568", "issue:44779", "issue:44855", "issue:44857", + "issue:44871", "issue:44977", - "issue:45072", "issue:45083", "issue:45229", "issue:45290", "issue:45381", "issue:45440", + "issue:45491", "issue:45588", "issue:45698", - "issue:45823", "issue:45920", "issue:46192", "issue:46399", @@ -9932,25 +9932,21 @@ "issue:46858" ], "soft_pairs": [ - "issue:44206|issue:45823", - "issue:32090|issue:33357", - "issue:42490|issue:43575", "issue:43550|issue:45083", - "issue:44857|issue:45072", "issue:44360|issue:44561", "issue:29942|issue:35141", - "issue:38175|issue:44493", "issue:27957|issue:34689", + "issue:38175|issue:44493", "issue:43278|issue:43329", - "issue:45920|issue:46192", "issue:29870|issue:32090", + "issue:45920|issue:46192", "issue:44448|issue:44977", "issue:45290|issue:45381", - "issue:10000|issue:15354", "issue:44977|issue:45229", + "issue:10000|issue:15354", "issue:43927|issue:45698", - "issue:40444|issue:43526", "issue:44855|issue:46399", + "issue:40444|issue:43526", "issue:43278|issue:43381", "issue:29870|issue:33290", "issue:45290|issue:46655", @@ -9960,210 +9956,221 @@ "issue:44561|issue:44779", "issue:46421|issue:46858", "issue:43873|issue:45229", - "issue:43976|issue:45440" + "issue:44448|issue:44466", + "issue:42490|issue:43873", + "issue:43976|issue:45440", + "issue:25251|issue:32090", + "issue:44568|issue:45491", + "issue:44871|issue:44977" ], - "prepared_review_unit_hash": "542a9ba9ceaa8a8c4fd28f30c49674f9790c0e36a45206df7d54e2440bd1a9a0", + "prepared_review_unit_hash": "79d984eb16b4a0dc11abbe077f7455b3e2bbb8f8d46a9ef31e9a6a9232395914", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36551, - "estimated_input_tokens": 9138, - "estimated_eval_tokens": 18532 + "soft_pair_count": 30, + "serialized_chars": 37064, + "estimated_input_tokens": 9266, + "estimated_eval_tokens": 18788 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most candidate pairs are unrelated bug reports from different subsystems and should not be merged as duplicates. If a single issue must act as a cluster anchor, the broad tokenizer regression report is the strongest general triage candidate, but it does not actually subsume the rest.", - "confidence": 0.58, - "canonical_issue_reason": "issue:45920 is the broadest high-signal regression in the set, with multiple affected models and inbound references, so it is the best general anchor if one must be chosen.", + "summary": "These are mostly unrelated bug reports across different subsystems and models. I did not find any soft pair that clearly describes the same underlying bug/change, so no duplicate edges are accepted. The broad tokenizer regression in issue 45920 is the best representative issue in this set.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45920 is the broadest and most central issue here: it describes a multi-model tokenizer regression in v5, matches several downstream symptoms, and has the strongest cluster-wide relevance.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 has the widest impact and clearest regression framing, making it the most suitable standalone issue to keep visible in this mixed cluster.", + "best_issue_reason": "issue:45920 is the best issue candidate because it captures a wider user-facing regression rather than a model-specific or narrow edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44206", - "right": "issue:45823", - "accept": false, - "reason": "Different failures in different code paths: LasrFeatureExtractor argument handling vs Gemma4 PLE device placement." - }, - { - "left": "issue:32090", - "right": "issue:33357", - "accept": false, - "reason": "Trainer broadcast NoneType error and MacOS CLIP bus error are unrelated symptoms and subsystems." - }, - { - "left": "issue:42490", - "right": "issue:43575", - "accept": false, - "reason": "Local path import warning and tensor-parallel OOM on Qwen2.5 are not the same bug." - }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile/SDPA issue in Bamba is unrelated to qwen3_omni_moe feature-length helper behavior." - }, - { - "left": "issue:44857", - "right": "issue:45072", - "accept": false, - "reason": "AMP/CUDA crash in LwDetr loss and bfloat16 dtype mismatch in SwitchTransformers/TimmWrapperModel are different problems." + "reason": "Different failures: torch.compile/SDPA crash vs a qwen3_omni_moe feature-length helper bug." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer ReLU discussion and trust_remote_code breakage from API removal are unrelated." + "reason": "Unrelated code paths: DSA indexer/ReLU discussion vs removal of is_torch_fx_available breaking trust_remote_code." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention test failures and post_init embedding reinitialization are separate model-loading/training issues." + "reason": "Flash Attention 2 test failures are not the same as output-embedding reinitialization during resize/post_init." }, { - "left": "issue:38175", - "right": "issue:44493", + "left": "issue:27957", + "right": "issue:34689", "accept": false, - "reason": "Siglip2 zero probabilities and unexpected key/position-id warnings are not the same underlying defect." + "reason": "One is XLMRoberta Flash Attention 2 support; the other is a Llama 3.2 vision model loading regression." }, { - "left": "issue:27957", - "right": "issue:34689", + "left": "issue:38175", + "right": "issue:44493", "accept": false, - "reason": "XLMRoberta Flash Attention 2 support and Llama 3.2 Vision model loading are distinct." + "reason": "Unexpected zero probabilities and unexpected position-id keys are different symptoms with no clear shared bug." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "BF16-to-FP32 evaluation dtype change is unrelated to undefined video-token helper variables." + "reason": "BF16-to-FP32 eval dtype drift is unrelated to undefined variables in the multimodal video token counter." }, { - "left": "issue:45920", - "right": "issue:46192", + "left": "issue:29870", + "right": "issue:32090", "accept": false, - "reason": "Both involve tokenizer APIs, but one is wrong token IDs across models while the other is a return_tensors keyword handling bug; not the same concrete issue." + "reason": "Generate-time decoder_attention_mask bug and Trainer GPU broadcast NoneType error are separate code paths." }, { - "left": "issue:29870", - "right": "issue:32090", + "left": "issue:45920", + "right": "issue:46192", "accept": false, - "reason": "BART generate() decoder_attention_mask bug and trainer broadcast NoneType error are unrelated." + "reason": "Tokenizer ID regression across models is not the same as return_tensors being ignored when text_kwargs is passed." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Pegasus v4/v5 output difference and Qwen3.5 flash-attention generation failure are different model-specific regressions." + "reason": "Pegasus v4/v5 output mismatch and Qwen3.5 flash-attention generation failure are different model-specific regressions." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Chat template/tool-call crash and Qwen2.5-VL video position-id bug are unrelated multimodal issues." + "reason": "Chat template crash on tool-call assistant messages is unrelated to video vision_position_ids being wrong." }, { - "left": "issue:10000", - "right": "issue:15354", + "left": "issue:44977", + "right": "issue:45229", "accept": false, - "reason": "German DistilBert issue and torch.jit GeneratorExp limitation on ViT export are unrelated." + "reason": "Flash-attention generation bug and multi-GPU CUDA OOM are different problems." }, { - "left": "issue:44977", - "right": "issue:45229", + "left": "issue:10000", + "right": "issue:15354", "accept": false, - "reason": "Flash-attention generation bug and multi-GPU CUDA OOM are different failure modes." + "reason": "Different model/export failures: DistilBert issue vs torch.jit.script generator-expression export limitation." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Custom token IDs lost after save/load and wrong custom module loaded after save_pretrained are related to persistence, but they are different concrete bugs." + "reason": "Both involve save/load, but one loses custom token IDs in config while the other loads the wrong custom module." }, { - "left": "issue:40444", - "right": "issue:43526", + "left": "issue:44855", + "right": "issue:46399", "accept": false, - "reason": "Qwen2.5-VL iterable dataset failure and BeitImageProcessorFast reduce_labels bug do not share the same code path." + "reason": "Python 3.13 IndentationError in DebertaV2Model is unrelated to a missing decorator in Qwen3.5 kernel registration." }, { - "left": "issue:44855", - "right": "issue:46399", + "left": "issue:40444", + "right": "issue:43526", "accept": false, - "reason": "Python 3.13 torch.jit parsing/import error and missing-decorator crash in Qwen3.5 kernels are separate issues." + "reason": "Multi-image iterable finetuning failure and BeitImageProcessorFast reduce_labels bug are unrelated." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Evaluation dtype mismatch and gradient checkpointing being disabled in eval mode are unrelated." + "reason": "Eval dtype change and gradient-checkpointing-in-eval limitation are distinct issues." }, { "left": "issue:29870", "right": "issue:33290", "accept": false, - "reason": "Generate-time decoder mask bug and deepspeed Adafactor OOM are unrelated." + "reason": "Generate() decoder attention mask bug is unrelated to DeepSpeed Adafactor OOM." }, { "left": "issue:45290", "right": "issue:46655", "accept": false, - "reason": "Chat-template tool-call crash and SmolVLM video padding bug are different multimodal failures." + "reason": "Chat-template tool-call crash and video padding miscalculation are different features and failures." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "Float16 AMP loss crash and flash_attention s_aux=None AttributeError are not the same bug." + "reason": "AMP/CUDA crash in LwDetrImageLoss is unrelated to flash_attention.py failing on s_aux=None." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer broadcast NoneType and deepspeed Adafactor OOM are unrelated." + "reason": "Trainer NoneType broadcast error and DeepSpeed optimizer OOM are different problems." }, { "left": "issue:37428", "right": "issue:42175", "accept": false, - "reason": "Flash-attention utility import error and missing TensorFlow backend in a torch install are not duplicates." + "reason": "Missing flash_attention_utils import and TensorFlow not included in the torch extra are unrelated." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Removal of is_torch_fx_available breaking trust_remote_code and DeepSeek tokenizer regression are different v5 regressions." + "reason": "Trust_remote_code breakage from removed helper is not the same as DeepSeek tokenizer regression." }, { "left": "issue:46421", "right": "issue:46858", "accept": false, - "reason": "Both concern cache/compile behavior, but one is spurious recompilation on chunked prefill while the other is a device-tensor seq-length arithmetic bug; not the same fix." + "reason": "Lazy cache initialization causing recompilation is different from get_seq_length returning a device tensor in static cache arithmetic." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Quantization offloading bug and Gemma4 multi-GPU OOM are distinct memory-management issues." + "reason": "Quantization offloading behavior and Gemma4 multi-GPU OOM are not the same bug." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility regression and DeepseekV3MoE divergence from remote implementation are unrelated." + "reason": "Python-version compatibility regression is unrelated to DeepSeekV3MoE diverging from the remote implementation." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Model output differences and lm_head.weight serialization inconsistency are separate issues." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Local-path import warning has no relation to quantization offloading behavior." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "top_k nesting in pipeline output is unrelated to the Trainer GPU broadcast NoneType error." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer BOS/EOS omission and Gemma3 NaN embeddings are different bugs." + }, + { + "left": "issue:43329", + "right": "issue:45381", + "accept": false, + "reason": "Undefined video helper variables and incorrect vision_position_ids are related to video processing broadly, but not clearly the same underlying bug." } ] }, @@ -10187,49 +10194,49 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-502", "nodes": [ - "issue:25251", "issue:29127", "issue:29870", "issue:29942", "issue:30064", - "issue:32090", "issue:33357", "issue:34567", "issue:34689", + "issue:35141", "issue:39692", "issue:41720", "issue:42490", "issue:42915", "issue:42994", + "issue:43065", "issue:43278", "issue:43381", + "issue:43412", "issue:43575", "issue:43611", "issue:43653", "issue:43701", "issue:43756", - "issue:43873", + "issue:43994", "issue:44186", "issue:44206", "issue:44265", "issue:44361", "issue:44442", - "issue:44448", - "issue:44466", "issue:44479", - "issue:44568", "issue:44743", - "issue:44871", "issue:44977", "issue:45198", "issue:45229", + "issue:45245", "issue:45372", "issue:45406", + "issue:45412", "issue:45440", "issue:45459", "issue:45491", "issue:45563", "issue:45588", + "issue:45593", "issue:45712", "issue:45803", "issue:45823", @@ -10237,16 +10244,11 @@ "issue:46693" ], "soft_pairs": [ - "issue:44448|issue:44466", - "issue:42490|issue:43873", - "issue:25251|issue:32090", - "issue:44568|issue:45491", - "issue:44871|issue:44977", - "issue:41720|issue:45229", "issue:29127|issue:34567", "issue:45588|issue:45823", - "issue:33357|issue:39692", + "issue:41720|issue:45229", "issue:43381|issue:43701", + "issue:33357|issue:39692", "issue:43278|issue:43575", "issue:44977|issue:45563", "issue:45198|issue:45803", @@ -10264,269 +10266,309 @@ "issue:45459|issue:45712", "issue:44265|issue:45198", "issue:45491|issue:46693", - "issue:45372|issue:45406" + "issue:45372|issue:45406", + "issue:43065|issue:43994", + "issue:43412|issue:45588", + "issue:29127|issue:33357", + "issue:45245|issue:45563", + "issue:43065|issue:45412", + "issue:43575|issue:43701", + "issue:45440|issue:45593", + "issue:42490|issue:43381", + "issue:45372|issue:45459", + "issue:29127|issue:35141" ], - "prepared_review_unit_hash": "c2f0156410065a23f22ee0fdf0e05e0860a182f0573939a77536ec9c640a38b0", + "prepared_review_unit_hash": "7fc93973aaaf53341920a088109cfffa4106e9fd65a3bb3fb66e8e0dcc10a8fb", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36373, - "estimated_input_tokens": 9094, - "estimated_eval_tokens": 18444 + "soft_pair_count": 33, + "serialized_chars": 37674, + "estimated_input_tokens": 9419, + "estimated_eval_tokens": 19094 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous set of closed/open issue reports rather than a tight duplicate cluster. The soft pairs are mostly near-misses across different models, symptoms, or code paths; none looks like the same underlying bug/change.", - "confidence": 0.88, - "canonical_issue_reason": "issue:43873 is the broadest, most active, and most generally framed report in the set (quantization/offloading behavior), while most other items are narrow, unrelated regressions or documentation/example failures.", + "summary": "This cluster is highly heterogeneous: most items are unrelated product/model-specific bugs, and none of the soft pairs look like true duplicates or the same fixable change. No soft edges should be merged.", + "confidence": 0.59, + "canonical_issue_reason": "If a single anchor is required, 46693 is the broadest and most current regression issue in the set (FA2/long-context generation), and it is the only item with any meaningful thematic proximity to another candidate. The cluster overall is not a true duplicate cluster, though.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the best representative issue because it is open, discussion-heavy, and describes a reusable underlying product bug rather than a one-off model-specific symptom.", + "best_issue_reason": "46693 is the least-specific, most central issue among otherwise unrelated reports, so it is the best representative anchor for triage. It still should not be treated as a duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44448", - "right": "issue:44466", - "accept": false, - "reason": "Both are v5 regressions, but one is output parity and the other is tied-weight serialization; different bugs." - }, - { - "left": "issue:42490", - "right": "issue:43873", - "accept": false, - "reason": "Trailing-slash import warning and quantization offloading behavior are unrelated." - }, - { - "left": "issue:25251", - "right": "issue:32090", - "accept": false, - "reason": "Pipeline top_k nesting and Trainer _gpu_broadcast_one TypeError are different code paths." - }, - { - "left": "issue:44568", - "right": "issue:45491", + "left": "issue:29127", + "right": "issue:34567", "accept": false, - "reason": "Tokenizer special-token handling and GPU NaNs from sliding-window attention are unrelated failures." + "reason": "Different models and failures: LayoutLMv3 error-message clarity vs TrainerState token-count update bug." }, { - "left": "issue:44871", - "right": "issue:44977", + "left": "issue:45588", + "right": "issue:45823", "accept": false, - "reason": "Gemma eos_token_id mismatch vs Qwen3.5 flash-attention generation; different models and symptoms." + "reason": "Different code paths: flash attention None handling vs Gemma4 PLE device placement mismatch." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Qwen3 auto device-map cuda assert and Gemma4 multi-GPU OOM are separate runtime issues." - }, - { - "left": "issue:29127", - "right": "issue:34567", - "accept": false, - "reason": "LayoutLMv3 error-message clarity and TrainerState token counting are unrelated." + "reason": "Both are OOM/compute failures, but on different models and causes; not the same bug." }, { - "left": "issue:45588", - "right": "issue:45823", + "left": "issue:43381", + "right": "issue:43701", "accept": false, - "reason": "Flash-attention s_aux=None crash and Gemma4 PLE device mismatch are different failures." + "reason": "Unrelated failures: eval-mode gradient checkpointing restriction vs checkpoint key mismatch." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "MacOS CLIP bus error and SigLIP2 docs/example errors are not the same bug." - }, - { - "left": "issue:43381", - "right": "issue:43701", - "accept": false, - "reason": "Eval-mode gradient checkpointing and resume checkpoint key mismatch are different issues." + "reason": "MacOS bus error on a community CLIP model vs SigLIP2 documentation/model-processor mismatch." }, { "left": "issue:43278", "right": "issue:43575", "accept": false, - "reason": "Embedding dtype drift in evaluate and tensor-parallel OOM are unrelated." + "reason": "Different issues: dtype regression in eval vs tensor-parallel load OOM." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "Flash-attention generation regression and stale warning text are different bug classes." + "reason": "Flash-attention generation bug vs a stale warning for num_return_sequences; not the same defect." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure and a bare-except error swallowing bug in Fuyu are unrelated." + "reason": "Wav2Vec2 save/tokenization failure vs a bare-except bug in Fuyu tensor conversion." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "LayoutLMv2 tokenizer crash and Wav2Vec2 save/tokenization fail involve different components." + "reason": "Different tokenizer/model families and different failure modes." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLukeTokenizer AttributeError is unrelated to the Wav2Vec2 failure." + "reason": "MLuke tokenizer AttributeError is unrelated to Wav2Vec2 save/tokenization failure." }, { "left": "issue:43611", "right": "issue:46121", "accept": false, - "reason": "base_model_prefix loading regression and rope-validation list handling are distinct code paths." + "reason": "Base-model-prefix loading regression vs rope-param validation TypeError; distinct code paths." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBirdTokenizer special-token registration and Qwen recurrent state reset are unrelated." + "reason": "BigBird tokenizer special-token registration bug vs Qwen recurrent-state reset bug." }, { "left": "issue:29870", "right": "issue:34689", "accept": false, - "reason": "BART generate() decoder mask bug and Llama 3.2 Vision loading breakage are different model-specific problems." + "reason": "Bart generate decoder mask bug vs Llama vision model loading break; no shared underlying fix." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma4 processor import dependency failure and flash_attention s_aux crash are separate issues." + "reason": "Gemma4 processor import failure vs flash_attention sink-less crash; different layers." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Smollm3 RoPE-layer mismatch and DeepSeekV3MoE divergence are unrelated model implementations." + "reason": "Smollm3 RoPE-layer mismatch vs DeepSeekV3MoE implementation divergence; not the same issue." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "FP8 config failure during load and quantized model saving failure are different stages of quantization." + "reason": "FineGrainedFP8 model-load failure vs quantized model saving failure." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Unsupported center arg in an extractor and Qwen video-input regression are unrelated." + "reason": "Unsupported feature-extractor arg vs video-input regression; different subsystems." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation-map processing are different subsystems." + "reason": "Flash Attention 2 test failures vs image processor void-segmentation handling." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBird tokenizer special-token bug and AutoTokenizer loading FastSpeech2ConformerTokenizer are unrelated." + "reason": "BigBird mask-token issue is unrelated to FastSpeech2ConformerTokenizer loading." }, { "left": "issue:45459", "right": "issue:45712", "accept": false, - "reason": "Hidden tokenizer errors from protobuf handling and leftover dummy classes are not the same bug." + "reason": "Exception handling for protobuf decode errors vs leftover dummy class cleanup; different fixes." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export compatibility and Wav2Vec2 save/tokenization failures are different problems." + "reason": "torch.export exportability issue vs Wav2Vec2 save/tokenization failure; not the same code path." }, { "left": "issue:45491", "right": "issue:46693", "accept": false, - "reason": "Gemma3 NaN embeddings and FA2 long-context performance regression are related only loosely by attention, not the same bug." + "reason": "Both involve FA2/long-context generation, but one is NaN correctness and the other is a performance regression; not the same bug." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Gemma4 processor import failure and serve-side missing _tokenizer attribute are different concrete issues." - } - ] - }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, - "error_kind": null, - "error_message": null - }, - { - "label": "issue", - "component_index": 1, - "component_count": 24, - "review_unit_index": 34, - "review_unit_count": 61, - "status": "reviewed", - "reason": null, - "source": "llm", - "cache_hit": false, - "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-502", - "nodes": [ - "issue:27957", - "issue:29127", - "issue:33290", - "issue:33357", - "issue:34689", - "issue:35141", - "issue:42490", + "reason": "Processor import error vs transformers serve missing _tokenizer attribute; unrelated failures." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Dummy Conv2d cleanup vs SigLIP2 nonsensical outputs; no shared underlying defect." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export guard/support issue vs flash_attention attribute crash; unrelated." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and CLIP bus error are unrelated." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count runtime limit vs paged generate warning; different concerns." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy class cleanup vs RT-DETR memory release bug; distinct issues." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM vs resume_from_checkpoint key mismatch; not the same change." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepSeek implementation divergence vs D-FINE auxiliary-loss behavior; unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Path-trailing-slash import warning vs gradient checkpointing eval-mode restriction." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma4 import failure and protobuf error masking are separate tokenizer/processor problems." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error messaging vs embedding reinitialization after resize; different functionality." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 34, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:33290", + "issue:34689", + "issue:35141", "issue:42915", - "issue:43065", "issue:43066", "issue:43334", "issue:43381", - "issue:43412", "issue:43421", "issue:43526", "issue:43531", "issue:43575", - "issue:43701", + "issue:43646", + "issue:43761", "issue:43824", "issue:43873", "issue:43931", "issue:43994", + "issue:44077", "issue:44186", "issue:44279", "issue:44387", "issue:44442", "issue:44462", "issue:44464", + "issue:44514", + "issue:44743", + "issue:44811", + "issue:44898", + "issue:44977", "issue:45072", "issue:45245", "issue:45250", - "issue:45372", + "issue:45381", "issue:45412", - "issue:45440", - "issue:45459", "issue:45491", - "issue:45563", "issue:45584", "issue:45588", - "issue:45593", "issue:45600", "issue:45834", + "issue:45859", "issue:46121", "issue:46369", "issue:46399", @@ -10534,259 +10576,231 @@ "issue:46693" ], "soft_pairs": [ - "issue:43065|issue:43994", - "issue:43412|issue:45588", - "issue:29127|issue:33357", - "issue:43065|issue:45412", - "issue:45245|issue:45563", - "issue:45440|issue:45593", - "issue:43575|issue:43701", - "issue:42490|issue:43381", - "issue:45372|issue:45459", - "issue:29127|issue:35141", - "issue:43421|issue:45491", "issue:43066|issue:44462", + "issue:43421|issue:45491", "issue:43994|issue:45412", - "issue:45412|issue:46369", "issue:43575|issue:44387", + "issue:45412|issue:46369", "issue:44442|issue:45072", "issue:43526|issue:45834", "issue:44464|issue:45584", "issue:42915|issue:43931", "issue:43526|issue:46693", - "issue:43994|issue:46369", "issue:43824|issue:44279", + "issue:43994|issue:46369", "issue:27957|issue:29127", "issue:27957|issue:33290", "issue:43526|issue:44186", "issue:33290|issue:34689", "issue:33290|issue:35141", "issue:46121|issue:46399", - "issue:45250|issue:46640", "issue:43334|issue:43531", + "issue:45250|issue:46640", + "issue:45588|issue:45600", "issue:43381|issue:43873", - "issue:45588|issue:45600" + "issue:43646|issue:44077", + "issue:43761|issue:45859", + "issue:44514|issue:45381", + "issue:44977|issue:45245", + "issue:44743|issue:44811", + "issue:44442|issue:44898" ], - "prepared_review_unit_hash": "de15c49489a91877015b6d37c798cb7627e6da4b8bf1ec0f9c670bce5fdfa25b", + "prepared_review_unit_hash": "bf58f0ed4c7f5bf4f5528a4ff2afc89bc6a76ca182a9675e5b768002dc6f956a", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 32, - "serialized_chars": 36432, - "estimated_input_tokens": 9108, - "estimated_eval_tokens": 18472 + "soft_pair_count": 28, + "serialized_chars": 35299, + "estimated_input_tokens": 8825, + "estimated_eval_tokens": 17906 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most pairs share only broad Transformer/modeling vocabulary, not the same underlying bug or change. No PRs are present. The only loose thematic overlap is a small Flash Attention 2 subgroup, but even there the titles point to different problems.", - "confidence": 0.95, - "canonical_issue_reason": "If a single issue must be chosen, issue:46640 is the broadest Flash Attention 2-related umbrella, but the collection is too mixed for it to be a true duplicate canonical.", + "summary": "This cluster is mostly a loose collection of unrelated Transformers issues with a few flash-attention/Qwen-adjacent items; none of the soft pairs look like true duplicates with the evidence shown.", + "confidence": 0.18, + "canonical_issue_reason": "issue:46640 is the best umbrella candidate: it is the most general flash-attention-2 report in the set and sits closest to the small FA2-related subgroup, though the overall cluster is too heterogeneous to treat as one duplicate family.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46640 is the most umbrella-like issue in the set, but it only fits a narrow Flash Attention 2 theme and does not represent the majority of items well.", + "best_issue_reason": "issue:46640 is the most representative issue for this set because it is broad, current, and closest to the only visible thematic thread (Flash Attention 2).", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43065", - "right": "issue:43994", - "accept": false, - "reason": "Different problems: Sam3PixelDecoder dummy Conv2d vs SigLIP2 bad outputs on AutoModel/pipeline." - }, - { - "left": "issue:43412", - "right": "issue:45588", - "accept": false, - "reason": "Different failure modes and components: executorch export guard/erfinv vs flash_attention s_aux=None crash." - }, - { - "left": "issue:29127", - "right": "issue:33357", - "accept": false, - "reason": "Unrelated: LayoutLMv3 box validation message clarity vs CLIP model bus error on macOS." - }, - { - "left": "issue:43065", - "right": "issue:45412", - "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d is not the same as RT-DETR memory not being released." - }, - { - "left": "issue:45245", - "right": "issue:45563", - "accept": false, - "reason": "Different bugs: category-limit RuntimeError vs stale num_return_sequences warning in paged generate()." - }, - { - "left": "issue:45440", - "right": "issue:45593", - "accept": false, - "reason": "Different models and behavior: DeepseekV3MoE divergence vs D-FINE auxiliary losses disabled." - }, - { - "left": "issue:43575", - "right": "issue:43701", - "accept": false, - "reason": "OOM while loading Qwen2-57B with tensor parallelism is unrelated to resume_from_checkpoint key mismatch." - }, - { - "left": "issue:42490", - "right": "issue:43381", - "accept": false, - "reason": "Local-path import warning has nothing to do with gradient checkpointing in eval mode." - }, - { - "left": "issue:45372", - "right": "issue:45459", - "accept": false, - "reason": "Different tokenizer/import issues: mistral_common ReasoningEffort import vs protobuf decode-error handling." - }, - { - "left": "issue:29127", - "right": "issue:35141", + "left": "issue:43066", + "right": "issue:44462", "accept": false, - "reason": "LayoutLMv3 error clarity is unrelated to token embedding resize reinitialization behavior." + "reason": "Both concern tokenizer loading behavior, but one is about decoder type in v5 and the other about ignoring tokenizer.json; different bugs and code paths." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Runtime special-token post-processor updates are unrelated to Gemma3 sliding-window NaN embeddings." - }, - { - "left": "issue:43066", - "right": "issue:44462", - "accept": false, - "reason": "Both mention tokenizer behavior, but one is wrong decoder type in v5 and the other ignores tokenizer.json; distinct issues." + "reason": "One is about runtime special-token/post-processor updates, the other about Gemma3 NaNs from sliding-window all-padding cases; unrelated failures." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 output quality and RT-DETR memory release are unrelated." + "reason": "AutoModel/pipeline wrong results for SigLIP2 is unrelated to RT-DETR memory not being released." }, { - "left": "issue:45412", - "right": "issue:46369", + "left": "issue:43575", + "right": "issue:44387", "accept": false, - "reason": "Both are memory-related at a high level, but one is GC/leak behavior in RT-DETR and the other is SAM3 LiteText memory overflow." + "reason": "Both mention OOM, but one is tensor-parallel model loading and the other is int4 quantization reserved-memory growth; not the same bug." }, { - "left": "issue:43575", - "right": "issue:44387", + "left": "issue:45412", + "right": "issue:46369", "accept": false, - "reason": "Both mention OOM, but one is tensor-parallel model load and the other is int4 quantization reserved-memory regression." + "reason": "Both are memory-related, but one is a GC/leak issue in RT-DETR and the other is model-call memory overflow for sam3-litetext; too different." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "FastSpeech2 tokenizer loading failure is unrelated to dtype mismatches in bfloat16 inference." + "reason": "FastSpeech2 tokenizer loading and bfloat16 dtype mismatches are unrelated." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is not the same as Kosmos2.5 long-OCR index error." + "reason": "BeitImageProcessorFast reduce_labels and Kosmos2.5 OCR index errors are different image-processing bugs." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is unrelated to Whisper empty-transcription failure after align_special_tokens." + "reason": "Compiled-forward chunked generation inconsistency is unrelated to Whisper empty-transcription handling after align_special_tokens." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to Qwen3-VL weight-shape mismatch on load." + "reason": "Qwen3Moe FP8 config failure and Qwen3-VL weight-shape mismatch are separate model/loading problems." }, { "left": "issue:43526", "right": "issue:46693", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to FA2 long-context performance regression." + "reason": "reduce_labels behavior in an image processor is unrelated to an FA2 long-context performance regression." }, { - "left": "issue:43994", - "right": "issue:46369", + "left": "issue:43824", + "right": "issue:44279", "accept": false, - "reason": "SigLIP2 pipeline output issue and SAM3 LiteText memory overflow are different model bugs." + "reason": "An import error for Qwen2_5_VL and a generic dependency issue are not the same concrete bug." }, { - "left": "issue:43824", - "right": "issue:44279", + "left": "issue:43994", + "right": "issue:46369", "accept": false, - "reason": "Missing Qwen2.5 VL import and generic dependency issue are not the same underlying problem." + "reason": "SigLIP2 bad outputs and sam3-litetext memory overflow do not share a code path." }, { "left": "issue:27957", "right": "issue:29127", "accept": false, - "reason": "XLMRoberta Flash Attention 2 support and LayoutLMv3 error-message clarity are unrelated." + "reason": "Flash Attention 2 support for XLMRoberta and LayoutLMv3 box-validation error messaging are unrelated." }, { "left": "issue:27957", "right": "issue:33290", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta is unrelated to Adafactor OOM under DeepSpeed." + "reason": "FA2 support for XLMRoberta and DeepSpeed AdaFactor OOM are different areas and causes." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug does not match LayoutLMv2Tokenizer NER/padding crashes." + "reason": "BeitImageProcessorFast label reduction and LayoutLMv2Tokenizer padding/NER crashes are separate tokenizer/processor bugs." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "Adafactor OOM in DeepSpeed is unrelated to Llama 3.2 Vision model loading breakage." + "reason": "AdaFactor/deepspeed OOM is unrelated to Llama 3.2 Vision model loading breakage." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "Optimizer OOM and token-embedding reinitialization are different code paths and failures." + "reason": "OOM in training with AdaFactor is not the same as embeddings being reinitialized during post_init." }, { "left": "issue:46121", "right": "issue:46399", "accept": false, - "reason": "RoPE validation TypeError with a JSON-loaded list is unrelated to the Qwen3.5 missing-decorator registration error." + "reason": "Both touch rope/rotary machinery, but one is a type-conversion validation issue and the other is a missing decorator registration crash." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id on Qwen3-VL text config and Qwen3-MoE sliding_window behavior are distinct issues." }, { "left": "issue:45250", "right": "issue:46640", "accept": false, - "reason": "Both mention Flash Attention 2, but one is a generic topic and the other is a specific issue title; insufficient evidence of the same bug/change." + "reason": "Both are about Flash Attention 2 in name only; one is a generic placeholder title and the other is a specific performance regression." }, { - "left": "issue:43334", - "right": "issue:43531", + "left": "issue:45588", + "right": "issue:45600", "accept": false, - "reason": "Qwen3-VL pad_token_id import failure is unrelated to Qwen3-MoE sliding_window behavior." + "reason": "Flash-attention AttributeError on sink-less models and removed Sam3LiteText auto-mapping refs are unrelated." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing in eval mode is unrelated to offloading not working with quantization." + "reason": "Gradient checkpointing in eval mode and quantization offloading behavior are different features and failures." }, { - "left": "issue:45588", - "right": "issue:45600", + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "A broad Transformers 5.0 custom initialization regression is not the same as patchtsmixer's optional post_init constraint." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen3_5Moe rotary embedding CPU-offload incompatibility are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Batched apply_chat_template padding=False crash and Qwen2.5-VL video vision_position_ids bug are different Qwen2.5-VL failure modes." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3.5 flash-attention generation issues and category-cardinality runtime limits are unrelated." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Qwen3.5 recurrent-state caching and Whisper batch_decode skip_special_tokens behavior are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:44898", "accept": false, - "reason": "Flash attention crash on s_aux=None is unrelated to auto_mappings.py referencing removed Sam3LiteText configs." + "reason": "FastSpeech2 tokenizer loading and Perceiver interpolation failures are different model/tokenizer bugs." } ] }, @@ -10814,12 +10828,15 @@ "issue:29870", "issue:42175", "issue:42490", + "issue:42907", + "issue:43232", "issue:43262", "issue:43316", + "issue:43388", "issue:43493", "issue:43525", "issue:43643", - "issue:43646", + "issue:43701", "issue:43720", "issue:43749", "issue:43761", @@ -10835,40 +10852,31 @@ "issue:44442", "issue:44479", "issue:44496", - "issue:44514", "issue:44743", - "issue:44811", "issue:44855", "issue:44871", - "issue:44898", "issue:44913", - "issue:44977", "issue:44987", "issue:45003", "issue:45137", "issue:45216", - "issue:45245", "issue:45325", - "issue:45381", "issue:45440", + "issue:45698", "issue:45759", "issue:45803", - "issue:45859", "issue:45907", "issue:46178", "issue:46234", - "issue:46697" + "issue:46506", + "issue:46697", + "issue:46829", + "issue:46932" ], "soft_pairs": [ - "issue:43646|issue:44077", - "issue:43761|issue:45859", - "issue:44514|issue:45381", - "issue:44977|issue:45245", - "issue:44743|issue:44811", - "issue:44442|issue:44898", "issue:43262|issue:44206", - "issue:45003|issue:45440", "issue:42490|issue:44743", + "issue:45003|issue:45440", "issue:43720|issue:44442", "issue:42175|issue:46234", "issue:43493|issue:43827", @@ -10884,207 +10892,199 @@ "issue:43316|issue:45759", "issue:44190|issue:44496", "issue:44442|issue:45803", - "issue:43761|issue:44188", "issue:43761|issue:44079", + "issue:43761|issue:44188", + "issue:44987|issue:46178", "issue:43931|issue:45216", - "issue:44987|issue:46178" + "issue:43232|issue:43388", + "issue:46829|issue:46932", + "issue:43701|issue:46506", + "issue:42907|issue:45698" ], - "prepared_review_unit_hash": "03565fd4918ed68c941265a1f79b63511beb2252a9f0cb17f74538ea1b3ec845", + "prepared_review_unit_hash": "e31b167131d4c1aeb53f09233e47e29cf50f6f3472aeb94e9a283c73ee66d527", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36376, - "estimated_input_tokens": 9094, - "estimated_eval_tokens": 18444 + "soft_pair_count": 26, + "serialized_chars": 35810, + "estimated_input_tokens": 8953, + "estimated_eval_tokens": 18162 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft edges should be rejected: the pairs only share broad model/family or loader/config themes, but each points to a different concrete bug or fix path.", - "confidence": 0.97, - "canonical_issue_reason": "No single canonical issue emerges here; the candidate pairs span unrelated regressions, config mismatches, loader bugs, docs issues, and model-specific behavior changes.", + "summary": "The cluster is highly heterogeneous; the soft pairs share at most broad subsystem vocabulary, not the same underlying bug. No PRs are present, and there is no clear canonical issue for the whole cluster.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No global best issue is appropriate because there is no coherent duplicate cluster to anchor a canonical triage target.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43646", - "right": "issue:44077", - "accept": false, - "reason": "Different problems: AutoConfig remote-code field loss vs patchtsmixer post_init policy; only both are config-related." - }, - { - "left": "issue:43761", - "right": "issue:45859", - "accept": false, - "reason": "Unrelated model paths: CLIPVision hidden_states regression vs Qwen rotary embedding CPU-offload incompatibility." - }, - { - "left": "issue:44514", - "right": "issue:45381", - "accept": false, - "reason": "Both touch Qwen2.5-VL, but one is batched chat-template padding and the other is video position-id scaling." - }, - { - "left": "issue:44977", - "right": "issue:45245", - "accept": false, - "reason": "Flash-attention generation bug is unrelated to the category-count runtime error." - }, - { - "left": "issue:44743", - "right": "issue:44811", - "accept": false, - "reason": "Different components and symptoms: recurrent cache reset vs Whisper batch_decode skip_special_tokens handling." - }, - { - "left": "issue:44442", - "right": "issue:44898", - "accept": false, - "reason": "Tokenizer loading failure and Perceiver image-size interpolation failure are distinct bugs." - }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and LASR feature-extractor center-arg crash are separate issues." + "reason": "Audio chat-template sampling-rate mismatch vs a LasrFeatureExtractor center-arg crash; unrelated code paths and failures." }, { - "left": "issue:45003", - "right": "issue:45440", + "left": "issue:42490", + "right": "issue:44743", "accept": false, - "reason": "sys.modules access safety and DeepSeekV3MoE divergence are unrelated code paths." + "reason": "Path-with-trailing-slash import warning is unrelated to recurrent-state reset in modular_qwen3_5 cache handling." }, { - "left": "issue:42490", - "right": "issue:44743", + "left": "issue:45003", + "right": "issue:45440", "accept": false, - "reason": "Local-path import warning is unrelated to recurrent-state cache reset behavior." + "reason": "Unsafe sys.modules access in modeling_utils is a different problem from DeepSeekV3MoE behavioral divergence." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "BitNet packed-weight loading and tokenizer startup failure do not describe the same bug." + "reason": "BitNet packed-weight loading failure and FastSpeech2ConformerTokenizer loading failure affect different loaders and models." }, { "left": "issue:42175", "right": "issue:46234", "accept": false, - "reason": "Packaging backend omission for tensorflow vs missing import_utils symbol; different dependency problems." + "reason": "Missing TensorFlow in a torch install extra is not the same as a missing import_utils availability helper." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation mismatch and docs still referencing pipeline() are unrelated." + "reason": "SigLIP2 implementation drift is unrelated to documentation still referencing removed pipelines." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "patchtsmixer post_init validation is unrelated to the Qwen video-input regression." + "reason": "patchtsmixer post_init allowance is a model-config issue, not the Qwen video-input regression." }, { "left": "issue:44062", "right": "issue:45907", "accept": false, - "reason": "AddedToken keyword collision and list_repo_files revision forwarding are different API bugs." + "reason": "AddedToken special-argument conflict and list_repo_files revision forwarding are unrelated API bugs." }, { "left": "issue:27957", "right": "issue:29870", "accept": false, - "reason": "XLMRoberta Flash Attention 2 support and BART decoder_attention_mask assignment error are distinct generation/model bugs." + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to Bart generate() using decoder_attention_mask before assignment." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Both involve config objects, but one is a Gemma3TextConfig API discrepancy and the other is a missing pad_token_id attribute." + "reason": "Gemma3TextConfig API discrepancy and missing pad_token_id on Llama4Config are different config-field problems." }, { "left": "issue:45325", "right": "issue:46697", "accept": false, - "reason": "Qwen2.5-VL rope indexing bug and Swin2SR over-padding are unrelated preprocessing/model behaviors." + "reason": "Qwen2.5-VL rope-index scaling bug and Swin2SR over-padding bug are different image/video preprocessing issues." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading bug and Gemma-3 eos_token_id config inconsistency do not match." + "reason": "Quantization/offloading behavior does not match the Gemma-3 eos_token_id configuration inconsistency." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "FSDP CPU RAM efficient loading regression and Qwen3-VL weight-shape mismatch are different loading failures." + "reason": "FSDP CPU RAM efficient loading and Qwen3-VL weight-shape mismatch are distinct loading failures." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 JIT parsing/IndentationError and DeepSpeed ZeRO3 deque underflow are unrelated." + "reason": "Python 3.13 decorator parsing error is unrelated to DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Remote-code AutoConfig field loss and GPTNeoX rotary_pct reload reset are different serialization/config issues." + "reason": "trust_remote_code field loss on AutoConfig reload is unrelated to GPTNeoX rotary_pct not persisting." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both concern composite configs, but one is a Gemma3TextConfig API mismatch and the other is AutoModelForCausalLM unwrapping text_config." + "reason": "Both mention config handling, but one is Gemma3TextConfig API shape and the other is AutoModelForCausalLM.from_config unwrapping text_config; not the same bug." }, { "left": "issue:44190", "right": "issue:44496", "accept": false, - "reason": "Dataset-loading script failure and unrecognized model/config error are unrelated." + "reason": "Local dataset loading in an example script is unrelated to an unrecognized model_type in config.json." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer load failure and FuyuBatchFeature exception swallowing are different bugs." + "reason": "Tokenizer loading failure and FuyuBatchFeature's bare except are different parts of the stack and different symptoms." }, { "left": "issue:43761", - "right": "issue:44188", + "right": "issue:44079", "accept": false, - "reason": "CLIPVision hidden_states regression and attention-kernel divergence under torch.compile are distinct model behaviors." + "reason": "CLIPVisionModel hidden_states regression is unrelated to ModelOutput key assignment when a key was previously None." }, { "left": "issue:43761", - "right": "issue:44079", + "right": "issue:44188", + "accept": false, + "reason": "Hidden_states regression in CLIPVisionModel and attention-kernel divergence under torch.compile are different bugs." + }, + { + "left": "issue:44987", + "right": "issue:46178", "accept": false, - "reason": "A vision-model forward regression is not the same as ModelOutput key assignment semantics." + "reason": "A model-loading failure for physical-intelligence/fast is unrelated to adding tensorflow_text to BACKENDS_MAPPING." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Weight-shape mismatch on load and incorrect save_pretrained checkpoint output are different Qwen3.5 issues." + "reason": "Qwen3-VL shape mismatch on load and Qwen3.5 save_pretrained checkpoint corruption are different lifecycle stages and causes." }, { - "left": "issue:44987", - "right": "issue:46178", + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs sync bug and gather_for_metrics truncating tuple labels are unrelated training/runtime issues." + }, + { + "left": "issue:46829", + "right": "issue:46932", + "accept": false, + "reason": "VideoPrism hidden_states/output bug is unrelated to Gemma4 quantized vision/audio embedder dtype casting." + }, + { + "left": "issue:43701", + "right": "issue:46506", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is unrelated to CPU/CUDA divergence in PPDocLayoutV3." + }, + { + "left": "issue:42907", + "right": "issue:45698", "accept": false, - "reason": "Loading a specific model under transformers>=5.1.0 and tensorflow_text backend registration are unrelated." + "reason": "Saving dequantized Ministrals/Devstrals and loading the wrong custom module after save_pretrained are different persistence issues." } ] }, @@ -11115,11 +11115,8 @@ "issue:28282", "issue:32090", "issue:42617", - "issue:42907", - "issue:43232", "issue:43295", "issue:43316", - "issue:43388", "issue:43404", "issue:43441", "issue:43452", @@ -11127,44 +11124,42 @@ "issue:43575", "issue:43606", "issue:43646", + "issue:43650", "issue:43701", "issue:43716", "issue:43742", + "issue:43761", "issue:43825", "issue:44077", "issue:44162", "issue:44242", "issue:44464", + "issue:44479", "issue:44496", "issue:44534", + "issue:44610", "issue:44792", "issue:44871", "issue:44913", "issue:45584", "issue:45632", - "issue:45698", "issue:45803", "issue:46050", "issue:46170", + "issue:46202", "issue:46234", "issue:46291", "issue:46365", + "issue:46369", "issue:46421", "issue:46424", - "issue:46506", "issue:46821", - "issue:46829", - "issue:46894", - "issue:46932" + "issue:46894" ], "soft_pairs": [ - "issue:43232|issue:43388", - "issue:42907|issue:45698", - "issue:46829|issue:46932", - "issue:43701|issue:46506", "issue:42617|issue:43575", - "issue:43550|issue:43701", "issue:43575|issue:43742", + "issue:43550|issue:43701", "issue:27957|issue:32090", "issue:43606|issue:43701", "issue:25251|issue:28282", @@ -11183,206 +11178,203 @@ "issue:45803|issue:46365", "issue:44162|issue:44792", "issue:43404|issue:43441", - "issue:44242|issue:46894", "issue:45632|issue:46424", - "issue:45803|issue:46170" + "issue:44242|issue:46894", + "issue:45803|issue:46170", + "issue:44610|issue:46369", + "issue:43650|issue:46202", + "issue:43761|issue:44479" ], - "prepared_review_unit_hash": "81b320b917517760bd2969e290402250fe74b1a5e975c2e4298afdbf10e56a5c", + "prepared_review_unit_hash": "02f2943582db37fb55673de99e02789487e6d5dcbcb6ce1e4a9956bbfcab558a", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 35928, - "estimated_input_tokens": 8982, - "estimated_eval_tokens": 18220 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 34962, + "estimated_input_tokens": 8741, + "estimated_eval_tokens": 17738 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The pairs are mostly loose topical matches rather than true duplicates; they point to different models, subsystems, or failure modes, so none should be merged as the same underlying bug/change.", - "confidence": 0.44, - "canonical_issue_reason": "Issue #45698 is the strongest canonical candidate: it is open, broadly scoped, has substantial discussion/inbound references, and describes a concrete save/load regression that is likely to attract duplicates.", + "summary": "This cluster is a loose set of unrelated Transformers issues with only superficial title overlap. None of the soft edges look like true duplicates or the same fixable code-path change, so no pair should be merged.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 44871 is the most concrete, self-contained regression report and the most central node among the soft links, but the surrounding items are heterogeneous rather than duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "#45698 is the best representative issue because it is actionable, user-facing, and broad enough to serve as the umbrella for related custom-module load/save reports.", + "best_issue_reason": "44871 is the best representative issue because it describes a specific, actionable bug with a narrow symptom; it is more triage-friendly than the more generic or broader reports.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43232", - "right": "issue:43388", - "accept": false, - "reason": "Different bugs: generation kwargs handling after sync_gpus vs metric gathering dropping tuple labels." - }, - { - "left": "issue:42907", - "right": "issue:45698", - "accept": false, - "reason": "Unrelated failures: a 3d_parallel script runtime issue vs custom module mismatch after save_pretrained/from_pretrained." - }, - { - "left": "issue:46829", - "right": "issue:46932", - "accept": false, - "reason": "Different model families and symptoms: VideoPrism output field mix-up vs Gemma4 quantized embedder dtype casting." - }, - { - "left": "issue:43701", - "right": "issue:46506", - "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to CPU/CUDA output divergence in PPDocLayoutV3." - }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "Both involve large/distributed workloads, but one is a 3d_parallel script failure and the other is an OOM on Qwen2 loading." + "reason": "Different failures: running a 3d_parallel example vs TP-loading Qwen2 causing OOM." }, { - "left": "issue:43550", - "right": "issue:43701", + "left": "issue:43575", + "right": "issue:43742", "accept": false, - "reason": "torch.compile/SDPA instability is a different code path from resume_from_checkpoint key mismatch." + "reason": "Different model-loading problems with different symptoms and likely different causes." }, { - "left": "issue:43575", - "right": "issue:43742", + "left": "issue:43550", + "right": "issue:43701", "accept": false, - "reason": "OOM on model parallel loading is not the same bug as a key error while loading MobileLLM-125M." + "reason": "Torch.compile/SDPA failure is unrelated to a resume_from_checkpoint key mismatch." }, { "left": "issue:27957", "right": "issue:32090", "accept": false, - "reason": "Flash Attention support for XLMRoberta and a NoneType broadcast error in Trainer are distinct issues." + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to a Trainer broadcast NoneType error." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU offload device mismatch is not the same as checkpoint key mismatch on resume." + "reason": "CPU offload device mismatch for bark-small is unrelated to checkpoint key naming." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline top_k output-shape change has nothing to do with missing PyTorch causing AutoModel import errors." + "reason": "Pipeline top_k output shape regression is unrelated to an ImportError about missing PyTorch." }, { "left": "issue:46234", "right": "issue:46291", "accept": false, - "reason": "One is a missing helper in import_utils; the other is an import failure from kernels' version/revision requirement." + "reason": "Missing utility symbol and kernels import-version failure are separate import-time issues." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Custom code accessing processor.tokenizer is a different regression from an incorrect pipeline translation support message." + "reason": "Processor/tokenizer regression is unrelated to an incorrect pipeline error message about translation tasks." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "A model_max_length question for DeBERTa-V3 is unrelated to a transformers.onnx import error." + "reason": "A model_max_length question and a missing transformers.onnx import are unrelated." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Processor/tokenizer access regression is unrelated to the Gemma-3 eos_token_id configuration mismatch." + "reason": "One is a processor/tokenizer API regression; the other is an eos_token_id config inconsistency." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Gemma-3 eos_token_id inconsistency is a different problem from Whisper empty-transcription generation failure." + "reason": "Gemma eos_token_id mismatch is unrelated to Whisper empty-transcription generation failure." }, { "left": "issue:44464", "right": "issue:46421", "accept": false, - "reason": "Both touch chunked generation/compiled paths, but one reports inconsistent outputs while the other reports spurious recompilation from cache lazy init." + "reason": "Chunked generation inconsistency with compiled forward is not the same as lazy cache initialization recompilation." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "Optional post_init allowance is not the same bug as non-persistent buffers being filled with junk in v5." + "reason": "PatchTSMixer post_init policy and non-persistent buffer corruption are different bugs." }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Gemma3TextConfig API discrepancy and Mistral-3 image preprocessor dtype mismatch are different model-specific issues." + "reason": "Gemma3TextConfig API discrepancy is unrelated to image preprocessor dtype mismatch." }, { "left": "issue:43316", "right": "issue:46821", "accept": false, - "reason": "A config API inconsistency is unrelated to TimesFM window_size raising AttributeError." + "reason": "Config API mismatch and TimesFM window_size AttributeError are different model-specific issues." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Gemma-3 eos token config mismatch is distinct from sinq quantized save/load breakage." + "reason": "Gemma eos_token_id inconsistency is unrelated to quantized sinq save/load breakage." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom model initialization breakage is not the same as GPTNeoX rotary_pct not persisting on reload." + "reason": "Custom model init breakage in Transformers 5 is not the same as rotary_pct not persisting on reload." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "gguf_file loading issues are different from an unrecognized model/config model_type error." + "reason": "gguf_file handling and unrecognized model config are different loading-path failures." }, { "left": "issue:45803", "right": "issue:46365", "accept": false, - "reason": "A bare except in Fuyu tensor conversion is unrelated to importing Trainer until datasets is uninstalled." + "reason": "Bare except in Fuyu tensor conversion is unrelated to the Trainer import issue involving datasets." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 being broken and a janus image-generation test failure are different models and failure modes." + "reason": "ESM2 breakage and a janus image-generation test failure are not the same bug." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Weight-tying in Mistral3ForConditionalGeneration is a different bug from Ministral-3 FlashAttention failures." + "reason": "lm_head weight tying in Mistral3 and FlashAttention failures in Ministral-3 are separate model regressions." }, { - "left": "issue:44242", - "right": "issue:46894", + "left": "issue:45632", + "right": "issue:46424", "accept": false, - "reason": "Missing load-balancing loss and a CohereASR double-shift training loss bug are not the same underlying change." + "reason": "Trust-remote-code cache path collisions are unrelated to max_cache_length being honored only for quantized cache." }, { - "left": "issue:45632", - "right": "issue:46424", + "left": "issue:44242", + "right": "issue:46894", "accept": false, - "reason": "Cache path collisions for trust_remote_code are unrelated to max_cache_length only applying to quantized cache." + "reason": "MoE load-balancing loss omission is not the same as a CohereASR double-shift training-loss bug." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "Swallowing exceptions in FuyuBatchFeature.convert_to_tensors is unrelated to Windows/ROCm torch.distributed import problems." + "reason": "Fuyu's bare except problem is unrelated to Windows/ROCm torch.distributed import dependency issues." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "OmDet-Turbo input-size mismatch and sam3-litetext memory overflow are different model problems." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "A vague data request is not a duplicate of a Gemma roundtrip failure." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVision hidden_states regression and video-input regression affect different code paths." } ] }, @@ -11418,21 +11410,18 @@ "issue:43531", "issue:43540", "issue:43572", - "issue:43650", "issue:43673", "issue:43688", - "issue:43761", "issue:43827", "issue:43994", "issue:44190", - "issue:44479", "issue:44514", "issue:44534", - "issue:44610", "issue:44617", "issue:44734", "issue:44743", "issue:44792", + "issue:44855", "issue:44869", "issue:44913", "issue:44964", @@ -11443,32 +11432,33 @@ "issue:45245", "issue:45406", "issue:45542", + "issue:45588", "issue:45600", + "issue:45632", "issue:45698", + "issue:46082", "issue:46178", - "issue:46202", "issue:46234", - "issue:46369", + "issue:46421", + "issue:46617", "issue:46640", + "issue:46697", "issue:46832", "issue:46858", "issue:46932" ], "soft_pairs": [ - "issue:44610|issue:46369", - "issue:43650|issue:46202", - "issue:43761|issue:44479", - "issue:43504|issue:45698", "issue:43540|issue:44514", + "issue:43504|issue:45698", "issue:45542|issue:46234", "issue:43388|issue:43688", - "issue:44792|issue:44977", "issue:43065|issue:44617", + "issue:44792|issue:44977", "issue:44792|issue:45600", - "issue:45127|issue:45200", "issue:43572|issue:46178", - "issue:43352|issue:46640", + "issue:45127|issue:45200", "issue:43673|issue:44743", + "issue:43352|issue:46640", "issue:41950|issue:43475", "issue:44190|issue:44964", "issue:44534|issue:44913", @@ -11478,180 +11468,190 @@ "issue:43827|issue:43994", "issue:45245|issue:46858", "issue:44869|issue:46832", - "issue:45072|issue:46932" + "issue:45072|issue:46932", + "issue:46082|issue:46697", + "issue:45632|issue:46617", + "issue:44855|issue:45588", + "issue:43673|issue:46421" ], - "prepared_review_unit_hash": "befd21b1d1feee70cf219978893103981b803cd98fe687f03ede59e7bd6248c2", + "prepared_review_unit_hash": "c73b1a833acc84cb07da6c6d4e64b98719f775ca1e5d50a3526bb7f5f7b9723a", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 24, - "serialized_chars": 34321, - "estimated_input_tokens": 8581, - "estimated_eval_tokens": 17418 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 25, + "serialized_chars": 35381, + "estimated_input_tokens": 8846, + "estimated_eval_tokens": 17948 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues span many unrelated subsystems and bug types; the suggested soft pairs are mostly keyword similarity rather than the same underlying defect, so none should be merged as duplicates.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "This cluster is heterogeneous: the soft pairs mostly share only broad subsystem vocabulary (cache, loading, flash-attention, processors) but not the same concrete bug or change. I would reject every soft edge and avoid collapsing these items into one duplicate cluster.", + "confidence": 0.98, + "canonical_issue_reason": "No strong canonical issue: the items span unrelated model, cache, processor, config, server, and docs problems, so there is no single duplicate target that represents the cluster.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "No global best issue; the set is not a true duplicate cluster, so any single issue would be an arbitrary representative rather than a meaningful canonical.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44610", - "right": "issue:46369", - "accept": false, - "reason": "Different failures: OmDet-Turbo input sizing vs SAM3 LiteText memory overflow; no shared code-path." - }, - { - "left": "issue:43650", - "right": "issue:46202", - "accept": false, - "reason": "Unrelated content: a placeholder issue versus a Gemma tokenizer/pipeline roundtrip bug." - }, - { - "left": "issue:43761", - "right": "issue:44479", + "left": "issue:43540", + "right": "issue:44514", "accept": false, - "reason": "Different models and regressions: CLIPVision hidden-states vs Qwen video-input behavior." + "reason": "Different bugs: one is a video-input processing failure in Qwen3OmniMoe, the other is a batched chat-template crash in Qwen2_5_VLProcessor with padding=False." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Different issues: legacy config field loading for BEIT vs wrong custom module resolution after save/load." - }, - { - "left": "issue:43540", - "right": "issue:44514", - "accept": false, - "reason": "Both touch multimodal Qwen, but one is a video-input ValueError and the other is chat-template batching; different code paths." + "reason": "Both concern loading, but one is a BEiT preset legacy-field failure and the other is a wrong custom-module reload after save_pretrained; different root causes." }, { "left": "issue:45542", "right": "issue:46234", "accept": false, - "reason": "Different backend/import problems: TensorFlow backend detection vs missing import_utils helper." + "reason": "Unrelated optional-dependency issues: tf backend ambiguity vs a missing import_utils availability helper for falcon mamba." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "Different training bugs: metric gathering label truncation vs auxiliary loss normalization." - }, - { - "left": "issue:44792", - "right": "issue:44977", - "accept": false, - "reason": "Different model-specific problems: janus image-generation test vs Qwen3.5 flash-attention generation." + "reason": "Evaluation label collation in gather_for_metrics is unrelated to auxiliary-loss normalization in OLMoE/GPT Oss." }, { "left": "issue:43065", "right": "issue:44617", "accept": false, - "reason": "Sam3PixelDecoder dummy layer concern is unrelated to Sam3Video OOM behavior." + "reason": "Same model family only superficially; a dummy Conv2d in Sam3PixelDecoder is not the same bug as Sam3Video CUDA OOM." }, { "left": "issue:44792", - "right": "issue:45600", + "right": "issue:44977", "accept": false, - "reason": "Unrelated: janus test failure vs auto_mappings stale config references." + "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation failure are different models and failure modes." }, { - "left": "issue:45127", - "right": "issue:45200", + "left": "issue:44792", + "right": "issue:45600", "accept": false, - "reason": "Both involve embeddings, but one is LoRA merge collapse and the other is missing multimodal token-type defaults." + "reason": "A Janus generation test failure is unrelated to stale auto_mappings references to removed Sam3LiteText configs." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "Separate configuration/import issues: StableLm pad token field vs tensorflow_text backend registration." + "reason": "StableLm pad_token_id regression and a tensorflow_text backend mapping error are different config/import problems." }, { - "left": "issue:43352", - "right": "issue:46640", + "left": "issue:45127", + "right": "issue:45200", "accept": false, - "reason": "Specific model unsupported Flash Attention 2 warning vs a generic flash-attention issue; not the same bug." + "reason": "LoRA merge collapse with tied embeddings is not the same as mm_token_type_ids defaulting for text-only fine-tuning." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both involve cache/state, but one is chunked prefill cache missing and the other is recurrent-state reset with seq_len>1." + "reason": "Both mention cache, but one is missing cache in chunked_prefill and the other is recurrent states resetting with seq_len>1; distinct concrete failures." + }, + { + "left": "issue:43352", + "right": "issue:46640", + "accept": false, + "reason": "One says NemotronHForCausalLM does not support Flash Attention 2.0 yet; the other is a separate generic flash-attention issue." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Different pipeline/model bugs: processor selection for video-classification vs missing encoder output attribute." + "reason": "Video-classification processor selection and a missing fpn_position_embeddings attribute in Sam3 vision output are unrelated." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Different loading failures: local dataset script issue vs multimodal model load regression." + "reason": "Local dataset loading in a training script is not the same bug as loading Phi-4-multimodal-instruct with the latest transformers." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Unrelated config/runtime bugs: junk non-persistent buffers vs rotary_pct reloading behavior." + "reason": "Non-persistent buffer junk initialization and GPTNeoX rotary_pct reload regression are unrelated configuration/runtime issues." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Different model config bugs: missing pad_token_id vs sliding_window handling in Qwen3-MoE." + "reason": "Missing pad_token_id in Llama4Config is unrelated to the Qwen3-MoE sliding_window issue." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Different code paths: serve KV-cache indexing crash vs Gemma4 processor missing tokenizer attribute." + "reason": "A serve KV-cache indexing crash and a Gemma4Processor missing _tokenizer crash are different server-side bugs." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Different concerns entirely: embedding reinitialization vs local-path import warning." + "reason": "Embedding reinitialization after resize and trailing-slash local-model import warnings do not describe the same problem." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "One is docs drift after pipeline removals; the other is a model correctness issue involving AutoModel/pipeline." + "reason": "Docs still referencing pipeline() after removals is not the same as SigLIP2 producing nonsense with AutoModel and pipeline." }, { "left": "issue:45245", "right": "issue:46858", "accept": false, - "reason": "Different runtime bugs: category-cardinality error vs static-cache slice arithmetic on non-CUDA backends." + "reason": "A category-count overflow and a device-tensor cache length arithmetic bug are unrelated." }, { "left": "issue:44869", "right": "issue:46832", "accept": false, - "reason": "Distinct model failures: Whisper timestamp decode edge case vs RT-DETR feature-level mismatch." + "reason": "Whisper timestamp decoding with trailing replacement characters is unrelated to RT-DETR feature-level mismatch in forward()." }, { "left": "issue:45072", "right": "issue:46932", "accept": false, - "reason": "Different dtype issues in different model families; not the same concrete bug." + "reason": "These are different dtype issues in different model paths: bfloat16 inference mismatches vs quantized vision/audio embedder casting." + }, + { + "left": "issue:46082", + "right": "issue:46697", + "accept": false, + "reason": "LlamaConfig head_dim validation and Swin2SR over-padding are unrelated bugs." + }, + { + "left": "issue:45632", + "right": "issue:46617", + "accept": false, + "reason": "Both involve trust_remote_code caching, but one is a leaf-directory cache collision and the other is a FileNotFoundError from cache loading; not the same root cause." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "An import/torch.jit parsing error in DebertaV2Model is unrelated to the flash_attention.py s_aux=None AttributeError." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Missing GenerationMixin cache during chunked_prefill and lazy cache initialization causing recompilation are distinct cache-related problems, not the same bug." } ] }, @@ -11684,20 +11684,22 @@ "issue:43065", "issue:43295", "issue:43335", + "issue:43526", "issue:43550", "issue:43644", - "issue:43673", "issue:43931", "issue:44060", "issue:44077", "issue:44112", "issue:44190", "issue:44242", + "issue:44462", "issue:44464", "issue:44479", "issue:44610", "issue:44617", "issue:44734", + "issue:44737", "issue:44779", "issue:44855", "issue:44877", @@ -11706,41 +11708,36 @@ "issue:45127", "issue:45245", "issue:45276", + "issue:45290", "issue:45357", "issue:45362", "issue:45381", "issue:45563", "issue:45584", - "issue:45588", "issue:45600", - "issue:45632", - "issue:46082", + "issue:45803", + "issue:45907", "issue:46369", - "issue:46421", "issue:46489", - "issue:46617", "issue:46655", "issue:46697", + "issue:46752", "issue:46897" ], "soft_pairs": [ - "issue:45632|issue:46617", - "issue:46082|issue:46697", - "issue:44855|issue:45588", - "issue:43673|issue:46421", - "issue:44242|issue:46897", "issue:43931|issue:44060", "issue:29870|issue:35141", "issue:44617|issue:45584", + "issue:44242|issue:46897", "issue:43644|issue:44913", "issue:43335|issue:44855", "issue:45563|issue:45584", - "issue:43550|issue:44190", "issue:25251|issue:27957", + "issue:43550|issue:44190", "issue:44933|issue:45362", "issue:44077|issue:44877", - "issue:45381|issue:46369", "issue:41720|issue:45245", + "issue:45381|issue:46369", "issue:44479|issue:44734", "issue:43335|issue:44190", "issue:43065|issue:44610", @@ -11751,211 +11748,225 @@ "issue:44779|issue:46489", "issue:44617|issue:46369", "issue:43295|issue:45276", - "issue:44112|issue:45600" + "issue:44112|issue:45600", + "issue:43526|issue:46697", + "issue:45290|issue:46752", + "issue:44462|issue:45907", + "issue:44737|issue:45803", + "issue:45276|issue:45584" ], - "prepared_review_unit_hash": "ab5f8e4bc3ad2702258f13b12af463d24749330c77a394b547fe69b283869d00", + "prepared_review_unit_hash": "3049808d6e407ea7a299d7545ab3e8e369a97ed161ca6808deec875746beffe9", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35589, - "estimated_input_tokens": 8898, - "estimated_eval_tokens": 18052 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36526, + "estimated_input_tokens": 9132, + "estimated_eval_tokens": 18520 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: most pairs are only loosely similar by subsystem or wording, not the same concrete bug/change. I would not merge any of the soft edges as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "These are mostly unrelated transformer/model bug reports that only share broad subsystem keywords. I found one plausible duplicate pair in the DeepSeek tokenizer regression on v5+; the rest are distinct issues and should stay separate.", + "confidence": 0.89, + "canonical_issue_reason": "issue:46489 is the clearest and most specific representative of the only true near-duplicate pair: it describes the DeepSeek tokenizer v5+ regression as a gap in the earlier fix, making it the best canonical issue for that bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43644 is the broadest, most centrally referenced regression in the set, but there is no strong duplicate cluster here.", + "best_issue_reason": "issue:46489 is the best overall issue here because it pinpoints a concrete, actionable tokenizer regression with an explicit tie to the prior fix path.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45632", - "right": "issue:46617", - "accept": false, - "reason": "Both involve trust_remote_code/cache loading, but one is a leaf-directory cache collision and the other is a FileNotFoundError; not the same bug." - }, - { - "left": "issue:46082", - "right": "issue:46697", - "accept": false, - "reason": "Unrelated models and failure modes: LlamaConfig head_dim validation vs image over-padding." - }, - { - "left": "issue:44855", - "right": "issue:45588", - "accept": false, - "reason": "Python 3.13 TorchScript parsing error is unrelated to flash_attention sink handling." - }, - { - "left": "issue:43673", - "right": "issue:46421", - "accept": false, - "reason": "Both mention chunked prefill/cache, but one is missing cache state and the other is spurious recompilation; different symptoms and likely different fixes." - }, - { - "left": "issue:44242", - "right": "issue:46897", - "accept": false, - "reason": "Different models and different training-loss bugs; only a vague pattern match." - }, { "left": "issue:43931", "right": "issue:44060", "accept": false, - "reason": "Both are Qwen3-family issues, but model shape loading and tied-weight warning are distinct code paths." + "reason": "Different models and different failures: Qwen3-VL weight-shape mismatch vs Qwen3-Next tied-weights warning." }, { "left": "issue:29870", "right": "issue:35141", "accept": false, - "reason": "Bart generate() decoder mask bug is unrelated to token-embedding resize/post_init reinitialization." + "reason": "Separate bugs: BART generate() hits an UnboundLocalError, while the other is output-embedding reinitialization after resize_token_embeddings." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription generation failure." + "reason": "Unrelated symptoms and code paths: Sam3Video CUDA OOM vs Whisper empty-transcription generation crash." + }, + { + "left": "issue:44242", + "right": "issue:46897", + "accept": false, + "reason": "Both concern training losses, but they are different model-specific bugs with different root causes." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Generic buffer initialization junk vs GPTNeoX config reload regression are different bugs." + "reason": "Different serialization/config problems: non-persistent buffers junk vs GPTNeoXConfig rotary_pct not persisting on reload." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "Sparse-layer creation in SwitchTransformers is unrelated to a Python 3.13 IndentationError." + "reason": "One is a SwitchTransformers config construction bug; the other is a Python 3.13 import/parsing failure in DebertaV2Model." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "Stale warning for num_return_sequences and Whisper empty-transcription crash are unrelated generation issues." + "reason": "Different generate() issues: a stale warning about num_return_sequences vs a crash on empty Whisper transcription." }, { - "left": "issue:43550", - "right": "issue:44190", + "left": "issue:25251", + "right": "issue:27957", "accept": false, - "reason": "torch.compile/SDPA model failure is unrelated to loading a local dataset in an example script." + "reason": "No shared bug: pipeline top_k nesting regression vs XLMRoberta Flash Attention 2 support issue." }, { - "left": "issue:25251", - "right": "issue:27957", + "left": "issue:43550", + "right": "issue:44190", "accept": false, - "reason": "Pipeline top_k nested-list behavior and XLMRoberta Flash Attention 2 are unrelated." + "reason": "Completely different failures: torch.compile/SDPA on Bamba vs local dataset loading in an example script." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Missing import from image_utils is unrelated to Qwen3.5 chat crashes." + "reason": "Unrelated: missing import from image_utils vs Qwen3.5-35B chat crash." }, { "left": "issue:44077", "right": "issue:44877", "accept": false, - "reason": "Optional post_init handling in patchtsmixer is unrelated to strict config loading for granite_speech." + "reason": "Both are config-related, but one is about patchtsmixer post_init allowance and the other about strict config loading for granite_speech." }, { - "left": "issue:45381", - "right": "issue:46369", + "left": "issue:41720", + "right": "issue:45245", "accept": false, - "reason": "Qwen2.5-VL video position IDs and Sam3-LiteText memory overflow are different problems." + "reason": "Different classes of bugs: Qwen3 auto device mapping cuda assert vs category-count limit RuntimeError." }, { - "left": "issue:41720", - "right": "issue:45245", + "left": "issue:45381", + "right": "issue:46369", "accept": false, - "reason": "Auto device mapping cudaErrorAssert and category-count limits are unrelated." + "reason": "Different issues: Qwen video-position IDs regression vs Sam3-LiteText memory overflow." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Video-input regression and serve KV-cache indexing crash do not share the same underlying bug." + "reason": "Both hit generation/video-serving paths, but the concrete bugs differ: video-input regression in Qwen models vs KV-cache continuation indexing in serve." }, { "left": "issue:43335", "right": "issue:44190", "accept": false, - "reason": "A SwitchTransformers config bug and an example script dataset-loading issue are unrelated." + "reason": "Unrelated bugs: sparse-layer creation in SwitchTransformers vs local dataset loading failure." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to OmDet-Turbo processor size mismatch." + "reason": "Different model/component bugs: a dummy Conv2d in Sam3PixelDecoder vs OmDet-Turbo processor/model size mismatch." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "SigLIP2 docs/model mismatch and local dataset loading failure are not the same issue." + "reason": "Docs/example errors in SigLIP2 are unrelated to the local dataset loading problem." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "LoRA merge/tied embeddings collapse and Qwen3.5 visual encoder save keys are distinct bugs." + "reason": "Different mechanisms: LoRA merge with extended vocab collapse vs save_pretrained writing wrong visual encoder keys." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Compiled forward inconsistency and a stale generate warning are different code-path problems." + "reason": "Different generation bugs: compiled-forward chunked generation inconsistency vs a stale warning for num_return_sequences." }, { "left": "issue:45381", "right": "issue:46655", "accept": false, - "reason": "Both are video-related regressions, but they affect different models and different preprocessing logic." + "reason": "Both are video-related, but the underlying bugs differ: wrong Qwen video position IDs vs over-padding in SmolVLM video preprocessing." }, { "left": "issue:44779", "right": "issue:46489", - "accept": false, - "reason": "Both concern DeepSeek tokenizer regressions, but the second is a specific gap in a prior fix and not clearly the exact same bug." + "accept": true, + "reason": "Same underlying bug: DeepSeek tokenizer output is wrong on transformers v5+, and 46489 is a follow-up describing the gap in the earlier fix." }, { "left": "issue:44617", "right": "issue:46369", "accept": false, - "reason": "CUDA OOM in Sam3Video and memory overflow in Sam3-LiteText are different model/runtime issues." + "reason": "Both mention Sam3 memory pressure, but they are different models and different reports (OOM vs memory overflow)." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Processor/tokenizer regression and gemma4 resize_token_embeddings behavior are unrelated." + "reason": "Distinct resize/tokenizer regressions with different affected code paths and symptoms." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "CI failures in GraniteSpeech and Sam3LiteText auto_mappings are separate breakages." + "reason": "Both are CI failures, but they stem from unrelated causes: stale GraniteSpeech device override test vs removed Sam3LiteText config references." + }, + { + "left": "issue:43526", + "right": "issue:46697", + "accept": false, + "reason": "Different image-processing bugs: BEiT label reduction vs Swin2SR over-padding." + }, + { + "left": "issue:45290", + "right": "issue:46752", + "accept": false, + "reason": "Both are chat-template crashes, but the triggers differ: assistant tool-call message with no content vs passing an empty message list." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Separate repository/loader bugs: AutoTokenizer ignoring tokenizer.json vs list_repo_files not forwarding revision." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "Unrelated: XLNet device placement bug vs FuyuBatchFeature swallowing exceptions." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Different domains entirely: Gemma4 embedding resize bug vs Whisper empty-transcription failure." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet. The DeepSeek tokenizer pair (issue:44779 / issue:46489) is the only clearly plausible duplicate-like match, and the other soft-edge verdicts are appropriately conservative given the titles and available evidence. I didn\u2019t see an overstatement of evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -11983,53 +11994,48 @@ "issue:41950", "issue:42175", "issue:42796", + "issue:43295", "issue:43404", - "issue:43526", "issue:43531", "issue:43540", + "issue:43582", "issue:43643", + "issue:43650", "issue:43673", "issue:43854", + "issue:43976", "issue:44016", "issue:44112", + "issue:44220", "issue:44246", "issue:44279", "issue:44295", - "issue:44462", "issue:44496", "issue:44521", "issue:44734", - "issue:44737", "issue:44938", "issue:44964", + "issue:44987", "issue:45003", + "issue:45083", "issue:45216", - "issue:45276", - "issue:45290", + "issue:45335", "issue:45381", "issue:45542", - "issue:45584", "issue:45600", "issue:45632", - "issue:45803", - "issue:45907", "issue:46082", "issue:46121", "issue:46234", + "issue:46424", "issue:46506", - "issue:46697", - "issue:46752", - "issue:46858" + "issue:46858", + "issue:46932" ], "soft_pairs": [ - "issue:45290|issue:46752", - "issue:43526|issue:46697", - "issue:44462|issue:45907", - "issue:44737|issue:45803", - "issue:45276|issue:45584", "issue:43540|issue:44734", - "issue:44521|issue:45381", "issue:44938|issue:45542", + "issue:44521|issue:45381", "issue:25251|issue:29870", "issue:43673|issue:46858", "issue:43854|issue:44964", @@ -12045,201 +12051,220 @@ "issue:36246|issue:41628", "issue:43404|issue:45216", "issue:32090|issue:35141", - "issue:43531|issue:45600", "issue:36246|issue:38617", + "issue:43531|issue:45600", "issue:41950|issue:44016", - "issue:44295|issue:46506" + "issue:44295|issue:46506", + "issue:43650|issue:44220", + "issue:43673|issue:46424", + "issue:43976|issue:44987", + "issue:43582|issue:46932", + "issue:43295|issue:45335", + "issue:45083|issue:46506", + "issue:44016|issue:44220" ], - "prepared_review_unit_hash": "4410de6f1876e9e08c036014f83ba999ad4b88a4b28b53c986571264a03bce8c", + "prepared_review_unit_hash": "6f538696449425b441e94c5620666effcb163fa47ec7ae9ed8df606f3d99a96b", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 36113, - "estimated_input_tokens": 9029, - "estimated_eval_tokens": 18314 + "soft_pair_count": 29, + "serialized_chars": 36493, + "estimated_input_tokens": 9124, + "estimated_eval_tokens": 18504 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated issues; the only loose thematic overlap is around `apply_chat_template`/multimodal input handling, but the soft pairs do not look like true duplicates. No PRs are present.", - "confidence": 0.34, - "canonical_issue_reason": "issue:44521 is the closest anchor for the only recurring theme here (`apply_chat_template` with multimodal inputs) and has the strongest discussion/inbound activity among the related-looking issues.", + "summary": "This cluster is highly heterogeneous: the soft-linked issues share broad themes like model loading, multimodal inputs, caching, or import errors, but they do not appear to describe the same underlying bug/change.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue stands out because the cluster spans many unrelated defects across different subsystems and models.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the most representative issue to keep at cluster level because it is broader, active, and connected to the shared chat-template/multimodal surface area.", + "best_issue_reason": "No issue is a strong representative for the whole cluster; the items are too mixed to anchor on one canonical bug.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45290", - "right": "issue:46752", - "accept": false, - "reason": "Both mention `apply_chat_template`, but one is about assistant tool-call messages with no content and the other is an empty-list `IndexError`; different failure modes and likely different fixes." - }, - { - "left": "issue:43526", - "right": "issue:46697", - "accept": false, - "reason": "Different image processors and different bugs (`reduce_labels` vs over-padding); only the broad preprocessing area overlaps." - }, - { - "left": "issue:44462", - "right": "issue:45907", - "accept": false, - "reason": "One is tokenizer loading behavior, the other is repo file listing/revision forwarding; not the same code path." - }, - { - "left": "issue:44737", - "right": "issue:45803", - "accept": false, - "reason": "XLNet device placement bug and Fuyu exception handling are unrelated." - }, - { - "left": "issue:45276", - "right": "issue:45584", - "accept": false, - "reason": "Separate model-specific resize-token-embeddings bug vs Whisper empty-transcription generation bug; no shared concrete defect." - }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "Different subsystems and failure mechanisms: Qwen3OmniMoe video input processing vs serve KV-cache continuation indexing." + "reason": "Different code paths: Qwen3OmniMoe video-input processing vs transformers serve KV-cache continuation indexing." }, { - "left": "issue:44521", - "right": "issue:45381", + "left": "issue:44938", + "right": "issue:45542", "accept": false, - "reason": "Both are multimodal-related, but one is assistant-mask generation in chat templates and the other is wrong `vision_position_ids` for Qwen2.5-VL video inputs; too different to merge." + "reason": "Unrelated failures: Python 3.14 import/load problem vs missing TensorFlow backend when only tensorboard is installed." }, { - "left": "issue:44938", - "right": "issue:45542", + "left": "issue:44521", + "right": "issue:45381", "accept": false, - "reason": "Python 3.14 import failure is unrelated to the TensorFlow backend packaging issue." + "reason": "Both are multimodal/Qwen-related, but one is assistant mask generation and the other is video position IDs; not the same bug." }, { "left": "issue:25251", "right": "issue:29870", "accept": false, - "reason": "Pipeline `top_k` output-shape regression and BART decoder attention mask bug are unrelated." + "reason": "Pipeline output nesting change and BART generate decoder_attention_mask error are unrelated." }, { "left": "issue:43673", "right": "issue:46858", "accept": false, - "reason": "Both involve cache internals, but one is a missing cache in v5 chunked prefill and the other is a device-tensor arithmetic bug in static cache under `torch.compile`; not the same fix." + "reason": "Both mention cache, but one is missing cache in chunked_prefill while the other is a device-tensor seq-length arithmetic issue." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Different model-loading failures for different models; no evidence of the same underlying bug." + "reason": "Both are model-load regressions, but they affect different models and likely different root causes." }, { "left": "issue:45632", "right": "issue:46234", "accept": false, - "reason": "`trust_remote_code` cache-path collision and missing import-utils availability helper are unrelated." + "reason": "Trust-remote-code cache-path collision and a missing import_utils symbol are unrelated." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Both are config-loading issues, but one loses fields with `trust_remote_code` while the other rejects a config without `model_type`; not the same defect." + "reason": "Both concern config/model loading, but one is missing returned fields with trust_remote_code and the other is an unrecognized model lacking model_type." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import latency and unsafe `sys.modules` access are not the same bug." + "reason": "Import-time performance complaint vs unsafe sys.modules access; not the same bug." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Distinct model-loading problems with different root causes and targets." + "reason": "Different problems: Phi-4 multimodal model loading vs stale auto_mappings references breaking CI." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "CI test flake in GraniteSpeech is unrelated to Phi-4 model loading." + "reason": "CI test flakiness in GraniteSpeech is unrelated to Phi-4 multimodal loading failure." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "General dependency/backend packaging complaint vs a separate dependency issue; no concrete shared code path." + "reason": "Dependency/backend packaging issue and a general dependency issue are too vague and not the same defect." }, { "left": "issue:27957", "right": "issue:35141", "accept": false, - "reason": "Flash Attention model support and token-embedding resize reinitialization are unrelated." + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to embedding reinitialization after resize_token_embeddings." }, { "left": "issue:46082", "right": "issue:46121", "accept": false, - "reason": "Both touch config validation, but one is `head_dim` acceptance and the other is rope-validation list handling; separate fixes." + "reason": "LlamaConfig head_dim validation and rope-param conversion TypeError are different configuration bugs." }, { "left": "issue:42796", "right": "issue:44521", "accept": false, - "reason": "Tokenizer BPE-marker handling and multimodal chat-template assistant masks are different surfaces." + "reason": "Tokenizer BPE-marker handling and multimodal assistant masks are unrelated." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Different missing-import symbols for different APIs; not the same underlying issue." + "reason": "Both are import errors, but they refer to different missing names and modules." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Weight tying in a generation model and checkpoint save correctness are different regression classes." + "reason": "Different regression types and models: untied lm_head in Mistral3ForConditionalGeneration vs incorrect saved checkpoint for Qwen3.5." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer broadcast error and embedding reinitialization bug are unrelated." + "reason": "Trainer broadcast TypeError and token-embedding resize reinitialization do not share a code path." }, { - "left": "issue:43531", - "right": "issue:45600", + "left": "issue:36246", + "right": "issue:38617", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and stale auto-mapping references are unrelated." + "reason": "Missing Qwen2_5_VLImageProcessor and missing layer_type_validation are unrelated import/export issues." }, { - "left": "issue:36246", - "right": "issue:38617", + "left": "issue:43531", + "right": "issue:45600", "accept": false, - "reason": "Missing exported names in different modules with no shared root cause." + "reason": "Qwen3-MoE sliding_window behavior and stale config references in auto_mappings are unrelated." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Pipeline image-processor lookup bug and notebook syntax error are unrelated." + "reason": "Video-classification pipeline bug and notebook syntax error are not duplicates." }, { "left": "issue:44295", "right": "issue:46506", "accept": false, - "reason": "Position-ids buffer handling and CPU/CUDA result divergence are different problems." + "reason": "Position_ids buffer access error and CPU/CUDA result divergence are different defects." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "Placeholder/data request issue and _torch_extract_fbank_features bug are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Both mention cache configuration, but they concern different semantics and code paths." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python-version compatibility failure and a model-loading failure for physical-intelligence/fast are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:46932", + "accept": false, + "reason": "AppleSilicon TypeError in allocator warmup and Gemma4 dtype-casting fix are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Processor/tokenizer regression and t5gemma decoder embedding resize bug are different issues." + }, + { + "left": "issue:45083", + "right": "issue:46506", + "accept": false, + "reason": "Feature-extractor length helper behavior in qwen3_omni_moe and PPDocLayoutV3 CPU/CUDA mismatch are unrelated." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and fbank feature extraction bug are not the same problem." } ] }, @@ -12271,28 +12296,28 @@ "issue:43441", "issue:43479", "issue:43526", - "issue:43582", - "issue:43650", + "issue:43577", "issue:43673", "issue:43824", "issue:43901", "issue:43976", "issue:44016", "issue:44062", - "issue:44220", + "issue:44075", "issue:44279", "issue:44327", "issue:44336", + "issue:44351", + "issue:44368", "issue:44496", "issue:44521", "issue:44683", "issue:44779", "issue:44792", "issue:44933", - "issue:44987", "issue:45042", - "issue:45083", "issue:45216", + "issue:45278", "issue:45335", "issue:45375", "issue:45405", @@ -12301,30 +12326,23 @@ "issue:45588", "issue:45593", "issue:45759", + "issue:45812", "issue:46082", "issue:46123", "issue:46369", "issue:46399", "issue:46424", - "issue:46506", + "issue:46512", "issue:46515", "issue:46693", - "issue:46752", - "issue:46932" + "issue:46752" ], "soft_pairs": [ - "issue:43650|issue:44220", - "issue:43673|issue:46424", - "issue:43976|issue:44987", - "issue:43582|issue:46932", - "issue:43295|issue:45335", - "issue:45083|issue:46506", - "issue:44016|issue:44220", "issue:44496|issue:45042", "issue:43526|issue:43901", - "issue:44683|issue:45446", - "issue:42907|issue:43441", "issue:43381|issue:45593", + "issue:42907|issue:43441", + "issue:44683|issue:45446", "issue:43901|issue:44496", "issue:43673|issue:44336", "issue:43479|issue:45375", @@ -12337,228 +12355,228 @@ "issue:44279|issue:46515", "issue:44521|issue:46693", "issue:45335|issue:45584", - "issue:46123|issue:46369", "issue:44327|issue:46752", + "issue:46123|issue:46369", "issue:43299|issue:45216", "issue:46082|issue:46399", "issue:44279|issue:45405", "issue:43295|issue:44779", - "issue:39692|issue:43901" + "issue:39692|issue:43901", + "issue:44075|issue:44368", + "issue:43577|issue:45216", + "issue:46082|issue:46512", + "issue:43976|issue:45405", + "issue:44779|issue:45812", + "issue:44351|issue:45278" ], - "prepared_review_unit_hash": "57438f2f92f9629fce07f612ead2341f59b17f33b58b125d7be32b4c9954767a", + "prepared_review_unit_hash": "0bd075336304f2516e0af688e7e2281a507787da8057cf7332192ed9ae39dae7", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36897, - "estimated_input_tokens": 9225, - "estimated_eval_tokens": 18706 + "soft_pair_count": 30, + "serialized_chars": 36769, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is not a true duplicate set: the items span unrelated bugs in configs, tokenizers, cache behavior, docs, and hardware/version-specific regressions. Similarity is mostly superficial (shared subsystem names or model families), not the same underlying defect.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue fits this cluster well; the reports are heterogeneous and do not describe one shared bug.", - "canonical_pr_reason": "No PRs are present in the cluster.", - "best_issue_reason": "issue:43295 is the clearest, most actionable report with a concrete regression and reproducible behavior, so it is the best representative if one must be chosen.", - "best_pr_reason": "No PRs are present, so there is no best PR candidate.", + "summary": "This cluster is highly heterogeneous: most items are isolated model/config/regression reports, so there is no strong duplicate core. I treated every soft edge as a false match because the paired artifacts do not describe the same concrete bug or change.", + "confidence": 0.63, + "canonical_issue_reason": "Issue 45278 is the broadest regression report in the set, describing multiple import failures after an upgrade. It is the closest thing to an umbrella issue, even though the cluster is mostly unrelated one-off bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the best stand-alone representative because it captures a wider API/import breakage rather than a single model-specific symptom.", + "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43650", - "right": "issue:44220", - "accept": false, - "reason": "Unrelated problems: one is an unclear data-related issue, the other is a specific fbank feature extraction bug." - }, - { - "left": "issue:43673", - "right": "issue:46424", - "accept": false, - "reason": "Both mention cache, but one is a GenerationMixin chunked-prefill cache regression and the other is cache length handling for quantized cache only." - }, - { - "left": "issue:43976", - "right": "issue:44987", - "accept": false, - "reason": "Different failure modes: Python version support vs loading a specific model checkpoint." - }, - { - "left": "issue:43582", - "right": "issue:46932", - "accept": false, - "reason": "Both touch platform-specific behavior, but one is an Apple Silicon allocator TypeError and the other is Gemma4 quantized embedder dtype casting." - }, - { - "left": "issue:43295", - "right": "issue:45335", - "accept": false, - "reason": "Different APIs and code paths: processor/tokenizer multimodal regression vs resize_token_embeddings not propagating to decoder embeddings." - }, - { - "left": "issue:45083", - "right": "issue:46506", - "accept": false, - "reason": "Different models and symptoms: helper length calculation in qwen3_omni_moe vs CPU/CUDA output mismatch in PPDocLayoutV3." - }, - { - "left": "issue:44016", - "right": "issue:44220", - "accept": false, - "reason": "A notebook syntax error is unrelated to a speech feature extraction bug." - }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "One is an unrecognized-model config problem; the other is an image-processor dependency regression." + "reason": "Different failures in different areas: model loading/config resolution vs PIL image processor torchvision dependency." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "One is a BEiT image processor label bug; the other is outdated pipeline docs text." + "reason": "One is a BeitImageProcessorFast label-processing bug; the other is outdated docs for text classification pipeline behavior." }, { - "left": "issue:44683", - "right": "issue:45446", + "left": "issue:43381", + "right": "issue:45593", "accept": false, - "reason": "Both are flex_attention-related, but they are different concrete bugs: compilation failure on torch>=2.9 vs a wrong version check for AuxRequest import." + "reason": "Gradient checkpointing eval-mode behavior and D-FINE denoising losses are unrelated code paths." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Different regressions: dequantized model saving vs FlashAttention failure during Ministral-3 loading." + "reason": "Both mention Ministral/devstral, but one is about saving dequantized models while the other is a FlashAttention load/runtime failure." }, { - "left": "issue:43381", - "right": "issue:45593", + "left": "issue:44683", + "right": "issue:45446", "accept": false, - "reason": "Unrelated model behaviors: gradient checkpointing in eval mode vs D-FINE auxiliary losses with denoising off." + "reason": "Both touch flex_attention, but one is a torch>=2.9 compiled-kernel failure and the other is a version-check/import bug for AuxRequest." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Docs drift in a text pipeline is not the same as a model-loading config error." + "reason": "Documentation mismatch for return_all_scores is unrelated to an unrecognized-model config error." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Cache-generation behavior and ANSI color handling are unrelated subsystems." + "reason": "Generation cache handling during chunked prefill is unrelated to ANSI coloring in loading_report." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Both are config-related, but one is default initialization when None is passed and the other is a missing field dropped by strict config handling." + "reason": "Both are config bugs, but they affect different multimodal model classes and different fields/defaulting behavior." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "A Python compatibility issue is not the same as a generic dependency issue." + "reason": "Python version support breakage is not the same underlying problem as a generic dependency issue." }, { "left": "issue:44496", "right": "issue:45759", "accept": false, - "reason": "Different model families and different code paths: unknown-model loading vs unwrapping text_config in AutoModelForCausalLM.from_config." + "reason": "Different multimodal model-construction failures: one rejects a model config, the other fails to unwrap text_config in from_config." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "A notebook syntax error is unrelated to the AddedToken keyword-argument bug." + "reason": "Notebook syntax error and AddedToken keyword conflict are unrelated." }, { "left": "issue:44792", "right": "issue:46424", "accept": false, - "reason": "A janus image-generation test failure is not the same as cache max_length semantics." + "reason": "A janus image-generation test failure is not the same as cache-length handling in quantized cache." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both involve flash_attention, but the concrete failures differ: Ministral-3 loading vs s_aux=None AttributeError in sink-less models." + "reason": "Both involve flash attention, but the concrete failures differ: Ministral-3 runtime breakage vs s_aux=None AttributeError for sink-less models." }, { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Both are import-related, but they concern different missing symbols and different modules." + "reason": "Missing Qwen2_5_VL export and a nonexistent image_utils import are separate import problems." }, { "left": "issue:44279", "right": "issue:46515", "accept": false, - "reason": "A dependency issue with transformers is not the same as a torch-version incompatibility." + "reason": "A generic transformers dependency issue is not the same as an explicit torch 2.4.1 incompatibility." }, { "left": "issue:44521", "right": "issue:46693", "accept": false, - "reason": "Different symptoms and scope: multimodal assistant masks vs FA2 long-context performance regression." + "reason": "Assistant-mask generation in multimodal chat templates is unrelated to FA2 long-context performance regression." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Tokenizer-embedding resizing bug and Whisper empty-transcription bug are unrelated." + "reason": "Embedding resize propagation in t5gemma is unrelated to Whisper empty-transcription generation after align_special_tokens." }, { - "left": "issue:46123", - "right": "issue:46369", + "left": "issue:44327", + "right": "issue:46752", "accept": false, - "reason": "A final-partial-batch flag bug is not the same as a memory overflow in sam3-litetext." + "reason": "QA span decoding top-k bounds and empty apply_chat_template input are different runtime errors." }, { - "left": "issue:44327", - "right": "issue:46752", + "left": "issue:46123", + "right": "issue:46369", "accept": false, - "reason": "Different edge cases in different pipelines: span decoding top-k bounds vs empty chat-template input." + "reason": "A partial-batch flag bug in MaskGenerationPipeline is unrelated to memory overflow in sam3-litetext." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Both concern Qwen multimodal models, but one breaks loading and the other breaks save_pretrained checkpoint output." + "reason": "Qwen3VL Moe loading regression and Qwen3.5 save_pretrained checkpoint corruption are different code-path failures." }, { "left": "issue:46082", "right": "issue:46399", "accept": false, - "reason": "Different model/config bugs: LlamaConfig head_dim validation vs a missing decorator for apply_rotary_pos_emb registration." + "reason": "LlamaConfig head_dim validation and Qwen3.5 kernel registration error are unrelated." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Unrelated dependency complaints: generic transformers dependency issue vs an unreleased minimum PEFT version bump." + "reason": "A vague dependency issue is not the same as an explicit MIN_PEFT_VERSION pin to an unreleased version." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Both mention v5 tokenizer/processor regressions, but they affect different model families and different behaviors." + "reason": "Custom model processor.tokenizer breakage and Deepseek tokenizer output regression are both tokenizer-related but not the same bug." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "Both are documentation-related, but one is a broken example with runtime errors and the other is stale parameter docs." + "reason": "SigLIP2 docs example errors are unrelated to a text-classification pipeline documentation mismatch." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Unused SGD optimizer args and a tie_word_embeddings warning are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation and Qwen3.5 save_pretrained checkpoint correctness are different regressions." + }, + { + "left": "issue:46082", + "right": "issue:46512", + "accept": false, + "reason": "Config validation for Llama head_dim has nothing to do with stale model.config.dtype after from_config(dtype=...)." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python support breakage and PEFT version pinning are separate packaging issues." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions affecting different model families and code paths." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A specific HybridCache import failure is not clearly the same underlying bug as the broader multi-import regression report." } ] }, @@ -12584,6 +12602,7 @@ "nodes": [ "issue:33290", "issue:41950", + "issue:42175", "issue:42796", "issue:42994", "issue:43299", @@ -12591,53 +12610,45 @@ "issue:43441", "issue:43479", "issue:43525", - "issue:43577", "issue:43650", "issue:43653", "issue:43854", "issue:43976", "issue:44016", - "issue:44075", "issue:44242", + "issue:44261", + "issue:44263", "issue:44291", "issue:44336", - "issue:44351", "issue:44361", - "issue:44368", "issue:44462", "issue:44466", + "issue:44623", + "issue:44671", "issue:44683", + "issue:44749", "issue:44779", "issue:45020", "issue:45081", "issue:45103", "issue:45200", - "issue:45216", "issue:45276", - "issue:45278", "issue:45357", "issue:45372", - "issue:45405", + "issue:45397", "issue:45542", "issue:45593", + "issue:45600", "issue:45656", - "issue:45812", + "issue:45712", "issue:45823", "issue:45834", "issue:45859", "issue:45907", - "issue:46082", "issue:46209", - "issue:46512", "issue:46851" ], "soft_pairs": [ - "issue:44075|issue:44368", - "issue:43577|issue:45216", - "issue:46082|issue:46512", - "issue:43976|issue:45405", - "issue:44779|issue:45812", - "issue:44351|issue:45278", "issue:42796|issue:43650", "issue:41950|issue:45020", "issue:43299|issue:45357", @@ -12658,198 +12669,204 @@ "issue:45103|issue:45372", "issue:44336|issue:46851", "issue:42994|issue:43317", - "issue:45859|issue:46851" + "issue:45859|issue:46851", + "issue:44623|issue:44749", + "issue:44261|issue:44263", + "issue:45600|issue:45712", + "issue:42175|issue:43976", + "issue:44671|issue:44779", + "issue:44623|issue:45397" ], - "prepared_review_unit_hash": "4de07989cdba4e21a4bc6fa3760d5f05887ace9ba86fc9d6ea988a7ffafe19ea", + "prepared_review_unit_hash": "49e96a24d55140ae976e95ec2309e3544ea8dd520c052768b4609a725ce34525", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35894, - "estimated_input_tokens": 8974, - "estimated_eval_tokens": 18204 + "serialized_chars": 35217, + "estimated_input_tokens": 8805, + "estimated_eval_tokens": 17866 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers bugs; despite superficial keyword overlap (tokenizers, loading, Qwen/Gemma/Deepspeed), none of the soft pairs look like the same concrete defect or mergeable fix.", - "confidence": 0.86, - "canonical_issue_reason": "Best representative is issue:44779 because it describes a broad v5 tokenizer regression with clear cross-model impact, but overall the cluster is too heterogeneous for a true duplicate canonical.", + "summary": "The proposed soft pairs are overwhelmingly false positives: they mostly span unrelated bug reports across tokenization, loading, quantization, DeepSpeed, model configs, and CI. There is no real duplicate cluster here.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue stands out; the items are heterogeneous and do not share one underlying bug or code path.", "canonical_pr_reason": null, - "best_issue_reason": "Issue:44779 is the most central, user-facing regression among the set, and it is the closest to a reusable duplicate target for nearby tokenizer-related reports.", + "best_issue_reason": "If one issue had to anchor the set, #44623 is the broadest standalone API regression, but it still does not represent the rest of the items.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44075", - "right": "issue:44368", - "accept": false, - "reason": "Both mention optimizer-related behavior, but one is about SGD args being ignored and the other about a LoRA warning in Qwen3.5; different code paths and fixes." - }, - { - "left": "issue:43577", - "right": "issue:45216", - "accept": false, - "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained checkpoint serialization are unrelated bugs." - }, - { - "left": "issue:46082", - "right": "issue:46512", - "accept": false, - "reason": "One is LlamaConfig validation for head_dim, the other is stale config.dtype after from_config; different symptoms and mechanisms." - }, - { - "left": "issue:43976", - "right": "issue:45405", - "accept": false, - "reason": "Python version compatibility and an unreleased MIN_PEFT_VERSION bump are not the same underlying issue." - }, - { - "left": "issue:44779", - "right": "issue:45812", - "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and appear to come from different tokenizer code paths." - }, - { - "left": "issue:44351", - "right": "issue:45278", - "accept": false, - "reason": "HybridCache import failure is a specific symbol-export problem; the other is a broad upgrade-related import breakage." - }, { "left": "issue:42796", "right": "issue:43650", "accept": false, - "reason": "Ministral tokenizer BPE-marker handling and a vague 'ADD THE DATA' issue are clearly unrelated." + "reason": "Tokenizer BPE-marker handling vs an unrelated data/title-only issue; no shared code path." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Video-classification processor lookup and remote_code regressions are different failures in different subsystems." + "reason": "Video pipeline processor lookup vs remote_code compatibility regressions; different subsystems." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Both involve Qwen3.x multimodal models, but one is MoE loading and the other is incorrect save_pretrained visual-encoder keys." + "reason": "MoE model loading failure vs save_pretrained visual-key serialization; different operations." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Tied-weight serialization inconsistency and Gemma4 embedding resize behavior are different model-save/model-resize bugs." + "reason": "Weight serialization inconsistency vs resize_token_embeddings propagation; not the same bug." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Qwen3VL MoE loading and GLM-4.7-Flash unit-test loading are unrelated model-specific failures." + "reason": "Different models and failure modes: Qwen3VL loading vs GLM-4.7 unit-test loading." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "Tokenizer special-token registration and D-FINE auxiliary-loss handling are unrelated." + "reason": "Tokenizer special-token registration vs D-FINE auxiliary-loss logic; unrelated." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "MoE load-balancing loss gating and Gemma4 mm_token_type_ids defaults are distinct functional bugs." + "reason": "Router/load-balancing loss computation vs mm_token_type_ids defaulting; different model paths." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Processor lookup regression and Gemma4 PLE device mismatch are not the same problem." + "reason": "Video processor lookup vs Gemma4 device-placement mismatch; no overlap." }, { "left": "issue:43650", "right": "issue:46209", "accept": false, - "reason": "A low-content placeholder issue cannot be the same as an FP8 inference failure on B200." + "reason": "Generic placeholder title vs FP8 inference on B200; no common bug." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Tokenizer.json being ignored and a Mistral regex patch crash are both tokenizer-related but not the same concrete defect." + "reason": "AutoTokenizer ignoring tokenizer.json vs Mistral regex patch crash; distinct tokenizer bugs." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Phi4MultimodalConfig default initialization and Llama4Config pad_token_id access are unrelated config bugs." + "reason": "Different config classes and attributes: Phi4Multimodal defaults vs Llama4 pad_token_id." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "A notebook syntax error and an init_empty_weights TypeError have no shared underlying code path." + "reason": "Notebook syntax error vs init_empty_weights constructor-argument mismatch." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and TensorBoard-without-TensorFlow backend errors are unrelated." + "reason": "torch flex_attention compilation bug vs missing TensorFlow backend dependency." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "MLukeTokenizer task-time AttributeError and D-FINE auxiliary-loss behavior are different subsystems." + "reason": "Tokenizer AttributeError on tasks vs D-FINE loss behavior; unrelated." }, { "left": "issue:45081", "right": "issue:45907", "accept": false, - "reason": "Mistral tokenizer patching and hub list_repo_files revision forwarding are unrelated API bugs." + "reason": "Tokenizer backend patch crash vs list_repo_files revision forwarding; different code paths." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "FlashAttention failure for Ministral-3 and Kosmos2.5 OCR index errors are different model-specific issues." + "reason": "FlashAttention/Mistral load failure vs Kosmos2.5 OCR index error; unrelated." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Both are Deepspeed/optimizer-related, but one is OOM with Adafactor while the other is double optimizer stepping." + "reason": "Adafactor OOM in DeepSpeed vs optimizer.step being called twice; not the same defect." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "auto_docstring annotations handling and a mistral_common import breakage are unrelated." + "reason": "auto_docstring annotation parsing vs Gemma4 processor import dependency." }, { "left": "issue:44336", "right": "issue:46851", "accept": false, - "reason": "ANSI escape leakage in loading_report and Qwen3.5 variable-length position_ids support are unrelated." + "reason": "ANSI escape cleanup vs variable-length position_ids handling." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized model saving failure and dequantized model loading with device_map=auto are distinct quantization pipeline bugs." + "reason": "Quantized model saving vs dequantized model loading with offload; different failures." }, { "left": "issue:45859", "right": "issue:46851", "accept": false, - "reason": "CPU-offload incompatibility in Qwen3_5MoeTextRotaryEmbedding and variable-length position_ids handling are different failures." + "reason": "CPU-offload rotary embedding incompatibility vs variable-length input handling." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Missing files in processor.save_pretrained vs a performance regression in data filtering." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "MLA q_a_layernorm epsilon config vs torch.split return values in GlmMoeDsaIndexer." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "auto_mappings stale config references vs leftover dummy classes in dummy_pt_objects." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Missing TensorFlow in extras install vs Python version support regression." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT masked-LM predictions vs DeepSeek tokenizer output regression." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "processor.save_pretrained issue vs ZeRO-3 from_pretrained; different code paths." } ] }, @@ -12874,10 +12891,13 @@ "cluster_id": "cluster-10000-502", "nodes": [ "issue:41628", - "issue:42175", + "issue:42491", "issue:42796", + "issue:43299", "issue:43441", "issue:43525", + "issue:43526", + "issue:43540", "issue:43643", "issue:43673", "issue:43723", @@ -12892,17 +12912,13 @@ "issue:44162", "issue:44188", "issue:44246", - "issue:44261", - "issue:44263", "issue:44297", "issue:44336", "issue:44373", "issue:44496", "issue:44589", - "issue:44623", - "issue:44671", - "issue:44749", - "issue:44779", + "issue:45003", + "issue:45072", "issue:45083", "issue:45103", "issue:45325", @@ -12911,23 +12927,19 @@ "issue:45542", "issue:45561", "issue:45563", - "issue:45600", - "issue:45712", "issue:45800", "issue:45823", "issue:46291", "issue:46399", + "issue:46614", "issue:46655", + "issue:46665", "issue:46726", + "issue:46752", + "issue:46829", "issue:46851" ], "soft_pairs": [ - "issue:44623|issue:44749", - "issue:44261|issue:44263", - "issue:45600|issue:45712", - "issue:42175|issue:43976", - "issue:44671|issue:44779", - "issue:44623|issue:45397", "issue:43673|issue:43761", "issue:43901|issue:43994", "issue:43901|issue:44246", @@ -12945,222 +12957,221 @@ "issue:45325|issue:46851", "issue:46655|issue:46726", "issue:41628|issue:43976", - "issue:45083|issue:46851", "issue:44246|issue:44496", + "issue:45083|issue:46851", "issue:43784|issue:45542", "issue:44297|issue:45397", "issue:46399|issue:46851", "issue:45561|issue:45563", - "issue:45103|issue:46399" + "issue:45103|issue:46399", + "issue:43526|issue:45003", + "issue:43540|issue:46752", + "issue:45072|issue:46829", + "issue:46614|issue:46665", + "issue:42491|issue:43299" ], - "prepared_review_unit_hash": "f1be90f6cd04375c587dccfa1f4dab1163dae67699323bbd5fe45fe0e66155a5", + "prepared_review_unit_hash": "7c1983e0a978bb23655a1b7acc358044bf13ae639202c8bc529ed669f45fa56b", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 36129, - "estimated_input_tokens": 9033, - "estimated_eval_tokens": 18322 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36457, + "estimated_input_tokens": 9115, + "estimated_eval_tokens": 18486 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All listed soft pairs look like adjacent-but-distinct issues rather than true duplicates. The strongest standalone issue is the Qwen2.5-VL import failure report, but there are no PRs in this cluster.", + "summary": "I found no safe duplicate merges in the soft pairs. Most matches are superficial keyword overlaps across different models, APIs, or regression types, not the same underlying bug.", "confidence": 0.93, - "canonical_issue_reason": "Issue 43824 is the highest-signal report: it names a precise missing export, has external references, and describes a concrete import failure path.", + "canonical_issue_reason": "No single canonical issue stands out: the cluster is heterogeneous, spanning unrelated subsystems and model families, so there is no clear duplicate target to anchor on.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43824 is the best representative issue overall because it is specific, actionable, and clearly tied to a reproducible code-path problem.", + "best_issue_reason": "Issue 43299 is the most concrete standalone regression in the set (a v5 loading breakage on the main branch), but it is still model-specific and does not represent the whole cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44623", - "right": "issue:44749", - "accept": false, - "reason": "Different problems: missing files on processor save vs a performance regression in data filtering after upgrade." - }, - { - "left": "issue:44261", - "right": "issue:44263", - "accept": false, - "reason": "Unrelated model-specific bug reports: one about missing `rms_norm_eps`, the other about `torch.split()` behavior in a different component." - }, - { - "left": "issue:45600", - "right": "issue:45712", - "accept": false, - "reason": "Both are repo-maintenance failures, but they target different files and failures: CI references vs dummy-class leakage." - }, - { - "left": "issue:42175", - "right": "issue:43976", - "accept": false, - "reason": "Packaging/backend install issue vs Python version compatibility regression; not the same underlying bug." - }, - { - "left": "issue:44671", - "right": "issue:44779", - "accept": false, - "reason": "Different model families and failure modes: CamemBERT masked LM predictions vs Deepseek tokenizer output." - }, - { - "left": "issue:44623", - "right": "issue:45397", - "accept": false, - "reason": "Processor save artifact problem vs `from_pretrained` load failure with ZeRO-3; no shared concrete code-path." - }, { "left": "issue:43673", "right": "issue:43761", "accept": false, - "reason": "Chunked prefill cache issue is unrelated to CLIPVisionModel hidden-states output behavior." + "reason": "Different regressions: cache handling in chunked prefill vs CLIP vision hidden_states output." }, { "left": "issue:43901", "right": "issue:43994", "accept": false, - "reason": "Docs mismatch for a pipeline flag vs incorrect model/pipeline results for SigLIP2; not the same bug." + "reason": "Docs mismatch for return_all_scores vs incorrect SigLIP model/pipeline outputs." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "A documentation inconsistency is unrelated to intermittent import latency." + "reason": "Documentation drift vs intermittent import slowness; not the same bug." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "Tokenizer loading regression and an ESM2 model breakage are different issues." + "reason": "Tokenizer loading issue vs ESM2 model breakage; unrelated failure modes." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "Missing Qwen2.5-VL export vs Python-version incompatibility; no shared fix." + "reason": "Missing imported model class vs Python version compatibility regression." }, { "left": "issue:42796", "right": "issue:43441", "accept": false, - "reason": "Tokenizer BPE marker handling and FlashAttention failure for Ministral are separate problems." + "reason": "Tokenizer BPE-marker handling vs FlashAttention failure in the same model family; different code paths." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Missing config attribute on Llama4 vs GLM-4V loading failure; different models and code paths." + "reason": "Llama4Config attribute error vs GLM-4V model loading failure; different models and causes." }, { "left": "issue:46291", "right": "issue:46399", "accept": false, - "reason": "Kernel import dependency/versioning issue vs missing decorator registration for rotary embedding; distinct failures." + "reason": "Kernel dependency/version check on import vs missing decorator in a Qwen3.5 kernel registration path." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "TensorFlow backend packaging issue vs torch/transformers version incompatibility; not the same bug." + "reason": "TensorBoard-without-TF backend error vs torch/transformers version incompatibility." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Torch.compile attention-kernel divergence is unrelated to ANSI output in loading_report." + "reason": "torch.compile attention divergence vs ANSI escape output in loading report; unrelated." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, - "reason": "Notebook syntax error and Gemma4 device mismatch are unrelated." + "reason": "Notebook syntax error vs device-map dtype/device mismatch in Gemma4 forward." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "A missing config field on one model is unrelated to `trust_remote_code` dropping fields in AutoConfig." + "reason": "Missing config attribute vs AutoConfig trust_remote_code field loss; different APIs." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "PyTorch storage object lookup failure is unrelated to Qwen3.5 chat crashes." + "reason": "Missing Float8 storage type vs Qwen3.5 chat crash; different code paths." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "Terminal-color output leakage is unrelated to a wrong docstring." + "reason": "Terminal coloring bug vs incorrect docstring; not duplicates." }, { "left": "issue:45325", "right": "issue:46851", "accept": false, - "reason": "Both involve position IDs, but they affect different models and different computations; too broad to treat as one bug." + "reason": "Different Qwen model families and different position_ids bugs (rope scaling vs variable-length handling)." }, { "left": "issue:46655", "right": "issue:46726", "accept": false, - "reason": "Video padding calculation bug vs missing-vision-backend image-processor error are different failure modes." + "reason": "Video processor padding bug vs missing image processor/backend detection." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "Missing `AutoImageProcessor` import and Python-version support issue are unrelated." - }, - { - "left": "issue:45083", - "right": "issue:46851", - "accept": false, - "reason": "A helper-length computation bug in qwen3_omni_moe is not the same as Qwen3.5 variable-length position_ids handling." + "reason": "Missing AutoImageProcessor export vs Python version support regression." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Import slowness and unrecognized model/config loading failure are unrelated." + "reason": "Import slowdowns vs unrecognized model/config loading failure." + }, + { + "left": "issue:45083", + "right": "issue:46851", + "accept": false, + "reason": "Helper output-length calculation in qwen3_omni_moe vs variable-length position_ids in Qwen3.5." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Sentence-transformers import NameError is unrelated to TensorFlow backend packaging." + "reason": "Import-time NameError in sentence-transformers integration vs TensorFlow backend detection issue." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer config metadata mismatch is unrelated to Gemma4 ZeRO-3 loading." + "reason": "tokenizer_config class naming mismatch vs zero3 from_pretrained failure." }, { "left": "issue:46399", "right": "issue:46851", "accept": false, - "reason": "Kernel registration crash and variable-length position_ids bug are different code paths." + "reason": "Kernel decorator registration crash vs position_ids variable-length bug; different subsystems." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "A pytest-xdist race in testing utils is unrelated to a stale warning in paged generate." + "reason": "xdist race in testing utils vs stale warning in paged generate." }, { "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "Auto-docstring crash under future annotations is unrelated to Qwen3.5 kernel registration." + "reason": "auto_docstring crash under future annotations vs Qwen3.5 kernel registration error." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "ImageProcessorFast label reduction bug vs unsafe sys.modules access in modeling_utils." + }, + { + "left": "issue:43540", + "right": "issue:46752", + "accept": false, + "reason": "Qwen3OmniMoe video processing ValueError vs empty-list chat template IndexError." + }, + { + "left": "issue:45072", + "right": "issue:46829", + "accept": false, + "reason": "bf16 dtype mismatch in model inference vs incorrect hidden_states returned by VideoPrism classifier." + }, + { + "left": "issue:46614", + "right": "issue:46665", + "accept": false, + "reason": "Stale phonemizer backend state vs inconsistent from_pretrained behavior in classes; not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen3 MoE loading, but one is a LoRA compatibility issue and the other is a model-loading regression; too broad to merge safely." } ] }, @@ -13186,13 +13197,12 @@ "nodes": [ "issue:36246", "issue:41950", - "issue:42491", "issue:42757", "issue:42886", "issue:43257", "issue:43262", - "issue:43299", - "issue:43526", + "issue:43295", + "issue:43475", "issue:43531", "issue:43540", "issue:43550", @@ -13208,10 +13218,10 @@ "issue:44479", "issue:44485", "issue:44496", + "issue:44749", "issue:44779", "issue:44871", "issue:45003", - "issue:45072", "issue:45137", "issue:45200", "issue:45292", @@ -13219,38 +13229,35 @@ "issue:45375", "issue:45397", "issue:45561", + "issue:45584", "issue:45588", "issue:45593", + "issue:45600", "issue:45759", + "issue:45859", "issue:46132", + "issue:46369", + "issue:46399", "issue:46506", - "issue:46614", "issue:46655", - "issue:46665", "issue:46752", "issue:46821", - "issue:46829", "issue:46851" ], "soft_pairs": [ - "issue:43526|issue:45003", - "issue:43540|issue:46752", - "issue:46614|issue:46665", - "issue:45072|issue:46829", - "issue:42491|issue:43299", "issue:43531|issue:45362", - "issue:45588|issue:46851", "issue:44246|issue:44871", - "issue:44242|issue:45593", + "issue:45588|issue:46851", "issue:44295|issue:45397", + "issue:44242|issue:45593", "issue:44336|issue:44485", "issue:45003|issue:46821", "issue:42886|issue:46132", "issue:43550|issue:46506", "issue:44451|issue:46752", "issue:44779|issue:45292", - "issue:43901|issue:45003", "issue:36246|issue:42757", + "issue:43901|issue:45003", "issue:43257|issue:45137", "issue:36246|issue:44351", "issue:41950|issue:44220", @@ -13258,206 +13265,218 @@ "issue:44496|issue:45003", "issue:43540|issue:46655", "issue:45375|issue:45759", - "issue:44162|issue:45200", "issue:43262|issue:44246", - "issue:44479|issue:46655" + "issue:44162|issue:45200", + "issue:44479|issue:46655", + "issue:45859|issue:46399", + "issue:44295|issue:44749", + "issue:43475|issue:46369", + "issue:45137|issue:45600", + "issue:43295|issue:45292", + "issue:45292|issue:45584" ], - "prepared_review_unit_hash": "f87c9b6dd9d2b0f8dcc91467bba4fe50aabd3a77e2fd85cf422e66bff333d356", + "prepared_review_unit_hash": "f881494e65a164bc7d58c7f2ee8716a9764ad41e64669ba831279ede070770a4", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35478, - "estimated_input_tokens": 8870, - "estimated_eval_tokens": 17996 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36306, + "estimated_input_tokens": 9077, + "estimated_eval_tokens": 18410 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: most pairs share only a broad subsystem or model family, not the same concrete bug. I found no safe duplicate merges among the soft edges.", + "summary": "These items are mostly unrelated Transformers bugs with only superficial overlap in model families or subsystems. I did not find any soft pair that clearly describes the same underlying bug/change, so all soft edges are rejected.", "confidence": 0.94, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue 44479 is the broadest and most representative regression here: it spans multiple Qwen multimodal/video variants and has a concrete version-specific failure mode, making it the best cluster anchor among these issues.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 44479 is the best representative issue because it has the widest blast radius and clearest regression framing, which makes it the most useful triage anchor in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43526", - "right": "issue:45003", - "accept": false, - "reason": "Different bugs: BeitImageProcessorFast label reduction vs unsafe sys.modules access in modeling_utils." - }, - { - "left": "issue:43540", - "right": "issue:46752", - "accept": false, - "reason": "Different code paths: Qwen3OmniMoe video processing vs empty-list apply_chat_template crash." - }, - { - "left": "issue:46614", - "right": "issue:46665", - "accept": false, - "reason": "Different failures: stale espeak backend state vs from_pretrained behavior inside Python classes." - }, - { - "left": "issue:45072", - "right": "issue:46829", - "accept": false, - "reason": "Both concern inference outputs, but one is a bf16 dtype mismatch and the other is wrong hidden_states assignment." - }, - { - "left": "issue:42491", - "right": "issue:43299", + "left": "issue:43531", + "right": "issue:45362", "accept": false, - "reason": "Same broad Qwen3-MoE area, but different variants and symptoms; not clearly one underlying loading bug." + "reason": "Different bugs: Qwen3-MoE sliding_window behavior vs a Qwen3.5 chat crash. Same family, different code paths and symptoms." }, { - "left": "issue:43531", - "right": "issue:45362", + "left": "issue:44246", + "right": "issue:44871", "accept": false, - "reason": "Qwen3-MoE sliding_window handling is not the same as Qwen3.5 chat crashes." + "reason": "Unrelated: a slow import problem versus a Gemma eos_token_id config mismatch." }, { "left": "issue:45588", "right": "issue:46851", "accept": false, - "reason": "Different issues in different paths: flash_attention sink handling vs variable-length position_ids." + "reason": "Both involve generation/input handling, but one is a flash-attention sinkless-model crash and the other is variable-length position_ids handling; not the same failure." }, { - "left": "issue:44246", - "right": "issue:44871", + "left": "issue:44295", + "right": "issue:45397", "accept": false, - "reason": "MoE load-balancing loss omission is unrelated to Gemma eos_token_id config mismatch." + "reason": "Different symptoms and areas: buffer registration access error versus a Gemma-4 ZeRO3 from_pretrained issue." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Both mention missing losses, but in different models and triggered by different flags." - }, - { - "left": "issue:44295", - "right": "issue:45397", - "accept": false, - "reason": "Different bugs: position_ids buffer access vs gemma-4 zero3 from_pretrained behavior." + "reason": "Both mention missing losses, but they affect different models and mechanisms; too broad to treat as one bug." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "ANSI-code output formatting is unrelated to GLM-5 RoPE implementation." + "reason": "ANSI/terminal output bug versus GLM-5 RoPE implementation; no shared underlying code-path." }, { "left": "issue:45003", "right": "issue:46821", "accept": false, - "reason": "Unsafe sys.modules access is not the same as the TimesFM window_size AttributeError." + "reason": "Modeling_utils sys.modules access bug is unrelated to the TimesFM window_size AttributeError." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Tokenizer offline-cache behavior differs from AttentionInterface.register changing function behavior." + "reason": "Offline tokenizer cache loading and AttentionInterface.register behavior change are different problems." }, { "left": "issue:43550", "right": "issue:46506", "accept": false, - "reason": "Different problems: Bamba torch.compile SDPA failure vs CPU/CUDA output drift in PPDocLayoutV3." + "reason": "Different runtime failures in different models/hardware paths; no evidence of the same bug." }, { "left": "issue:44451", "right": "issue:46752", "accept": false, - "reason": "Model load failure for ScandiBERT is unrelated to empty-list chat templating." + "reason": "Loading a specific model from cache is unrelated to apply_chat_template crashing on an empty list." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "DeepSeek tokenizer regression and resize_token_embeddings/output_embeddings mismatch are distinct." + "reason": "Tokenizer output regression and resize_token_embeddings not updating output embeddings are distinct issues." }, { - "left": "issue:43901", - "right": "issue:45003", + "left": "issue:36246", + "right": "issue:42757", "accept": false, - "reason": "Docs mismatch for return_all_scores is not the same as sys.modules access." + "reason": "Both are import errors, but for different symbols, packages, and contexts." }, { - "left": "issue:36246", - "right": "issue:42757", + "left": "issue:43901", + "right": "issue:45003", "accept": false, - "reason": "Both are import errors, but for different symbols and different dependencies." + "reason": "Docs/behavior mismatch in TextClassificationPipeline is unrelated to unsafe sys.modules access." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MOE weight conversion under accelerate+deepspeed is not the same as a DeepSpeed ZeRO3 deque IndexError." + "reason": "Both mention DeepSpeed, but one is about Qwen3 MOE weight conversion and the other is a ZeRO3 deque error." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Importing Qwen2_5_VLImageProcessor and HybridCache are separate missing-export problems." + "reason": "Import errors for different APIs; not the same missing-export bug." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification image-processor lookup and audio fbank extraction are unrelated." + "reason": "Video pipeline choosing image processors and _torch_extract_fbank_features() are different audio/video issues." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe sys.modules access and pytest-xdist file races are different concurrency/loading bugs." + "reason": "sys.modules access in modeling_utils is unrelated to pytest-xdist capture-file races." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model config is unrelated to sys.modules access." + "reason": "Unrecognized model/config loading is a different failure mode than unsafe sys.modules access." }, { "left": "issue:43540", "right": "issue:46655", "accept": false, - "reason": "Both are video-related, but Qwen3OmniMoe input validation and SmolVLM padding math are different bugs." + "reason": "Both are video-related, but they affect different models and the described causes differ (video input validation vs padding calculation)." }, { "left": "issue:45375", "right": "issue:45759", "accept": false, - "reason": "Both touch composite Qwen configs, but one is a missing field being dropped and the other is text_config unwrapping in from_config." + "reason": "Both concern Qwen multimodal config handling, but one is a missing config field and the other is AutoModelForCausalLM.from_config unwrapping text_config; not the same bug." }, { - "left": "issue:44162", - "right": "issue:45200", + "left": "issue:43262", + "right": "issue:44246", "accept": false, - "reason": "ESM2 breakage is unrelated to Gemma 4 mm_token_type_ids defaults." + "reason": "Audio chat-template sampling-rate behavior is unrelated to import performance." }, { - "left": "issue:43262", - "right": "issue:44246", + "left": "issue:44162", + "right": "issue:45200", "accept": false, - "reason": "Audio chat-template sampling-rate default and MoE load-balancing loss are different features." + "reason": "ESM2 breakage and Gemma-4 mm_token_type_ids defaults are unrelated model-specific bugs." }, { "left": "issue:44479", "right": "issue:46655", "accept": false, - "reason": "Both mention video inputs, but they affect different models and different failure mechanisms." + "reason": "Both mention video regressions, but they are for different model families and likely different processor logic." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "Related to Qwen3.5 rotary embedding code, but the reported failures are different enough that they do not look like one concrete bug." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "An attribute error around position_ids is unrelated to a performance regression after upgrading Transformers." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "SAM 3 Video encoder attribute loss and SAM3LiteText memory overflow are different model-specific failures." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque failure and broken auto_mappings references are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Processor.tokenizer regression and resize_token_embeddings/output_embeddings mismatch are different API bugs." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resizing and Whisper empty-transcription handling have no shared underlying code-path." } ] }, @@ -13486,56 +13505,50 @@ "issue:36246", "issue:42199", "issue:42886", - "issue:43295", + "issue:43065", "issue:43352", + "issue:43408", "issue:43475", "issue:43526", "issue:43540", "issue:43596", "issue:43653", + "issue:43723", "issue:44038", "issue:44060", "issue:44077", "issue:44162", "issue:44242", - "issue:44295", "issue:44368", "issue:44466", "issue:44479", "issue:44623", - "issue:44749", "issue:44857", "issue:44877", "issue:44977", "issue:44987", - "issue:45137", "issue:45229", "issue:45250", - "issue:45292", "issue:45335", "issue:45375", + "issue:45397", "issue:45412", "issue:45459", - "issue:45584", - "issue:45600", - "issue:45859", + "issue:45563", + "issue:45656", + "issue:45987", + "issue:46082", "issue:46123", "issue:46170", - "issue:46369", - "issue:46399", "issue:46506", + "issue:46710", "issue:46726", + "issue:46832", "issue:46851", "issue:46894", "issue:46897" ], "soft_pairs": [ - "issue:45859|issue:46399", - "issue:44295|issue:44749", - "issue:43475|issue:46369", - "issue:45137|issue:45600", - "issue:45292|issue:45584", - "issue:43295|issue:45292", "issue:44877|issue:45375", "issue:44466|issue:45335", "issue:36246|issue:43596", @@ -13546,210 +13559,216 @@ "issue:44038|issue:46506", "issue:29870|issue:30333", "issue:42199|issue:46851", - "issue:45412|issue:46123", - "issue:43653|issue:44162", - "issue:43526|issue:46894", "issue:44060|issue:44077", + "issue:45412|issue:46123", "issue:44479|issue:46170", - "issue:43352|issue:45250", + "issue:43526|issue:46894", + "issue:43653|issue:44162", "issue:44368|issue:44977", + "issue:43352|issue:45250", "issue:42886|issue:45229", "issue:44877|issue:46170", - "issue:43475|issue:43540" + "issue:43475|issue:43540", + "issue:45397|issue:45656", + "issue:46170|issue:46832", + "issue:43065|issue:43408", + "issue:43723|issue:46710", + "issue:44368|issue:45563", + "issue:45987|issue:46082" ], - "prepared_review_unit_hash": "94206f4a7cad7077dc4824d5bb91fcb3a2421d60f66a337c3a4055eed1f4b2d7", + "prepared_review_unit_hash": "f4142b2fbaf59ca89e479ac8a40e8bc90062fba75532495bc75a8d9d4c6aaae9", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 35165, - "estimated_input_tokens": 8792, - "estimated_eval_tokens": 17840 + "serialized_chars": 35073, + "estimated_input_tokens": 8769, + "estimated_eval_tokens": 17794 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is largely heterogeneous: the soft pairs share broad subsystem vocabulary (Qwen, strict config, embeddings, tokenizer, etc.) but not the same concrete bug or change. I would not merge any of the proposed soft edges.", - "confidence": 0.89, - "canonical_issue_reason": "No pair looks like a true duplicate; if a single issue must anchor the cluster, issue 44479 is the strongest representative because it describes a broad, version-specific regression affecting multiple Qwen video-input variants with clear user impact.", + "summary": "This set is not a single duplicate cluster; it contains many unrelated issues across tokenizers, vision/video models, config loading, training, and runtime behavior. The soft links are mostly superficial thematic matches, not the same underlying bug.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this heterogeneous set; the items describe distinct bugs in different subsystems and model families.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 is the most representative global issue: it is concrete, impactful, and scoped as a regression across several closely related models rather than a narrow edge case or auxiliary failure mode.", + "best_issue_reason": "No global best issue: none of these is representative of the whole set, and several only share broad surface symptoms rather than a common code path.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45859", - "right": "issue:46399", - "accept": false, - "reason": "Both mention Qwen3.5 rotary/attention internals, but one is CPU-offload compatibility for text rotary embedding and the other is a kernel registration error; different failure paths." - }, - { - "left": "issue:44295", - "right": "issue:44749", - "accept": false, - "reason": "These are unrelated: one is a position_ids buffer access error, the other is a general performance slowdown after upgrading Transformers." - }, - { - "left": "issue:43475", - "right": "issue:46369", - "accept": false, - "reason": "One is a missing attribute on Sam3VisionEncoderOutput; the other is a memory overflow on sam3-litetext. Different bugs and code paths." - }, - { - "left": "issue:45137", - "right": "issue:45600", - "accept": false, - "reason": "DeepSpeed ZeRO3 deque exhaustion is unrelated to auto_mappings referencing removed Sam3LiteText configs." - }, - { - "left": "issue:45292", - "right": "issue:45584", - "accept": false, - "reason": "resize_token_embeddings not updating output embeddings is a model-weight sync issue; Whisper empty-transcription failure is a generation/token alignment bug." - }, - { - "left": "issue:43295", - "right": "issue:45292", - "accept": false, - "reason": "Custom processor/tokenizer regression and token-embedding resizing are different problems with no shared code-path evidence." - }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Both involve config strictness, but one blocks loading granite_speech while the other silently drops a missing field in Qwen3_5MoeVisionConfig; not the same bug." + "reason": "Both involve strict config handling, but one is about loading granite_speech and the other about a missing Qwen3.5MoeVisionConfig field; different model/config paths." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Both touch tied/resize embedding behavior, but one is inconsistent serialization of lm_head.weight and the other is resize_token_embeddings not propagating to decoder.embed_tokens in T5Gemma." + "reason": "Both touch tied/embedding weights, but one is inconsistent lm_head serialization by device and the other is resize_token_embeddings not updating decoder.embed_tokens." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "An import error for Qwen2_5_VLImageProcessor and a DeepSpeed/BertModel index error are unrelated." + "reason": "Completely different failures: missing Qwen2.5 VL image processor import vs DeepSpeed ZeRO-3/BertModel index error." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Load-balancing loss omission in MoE routing is not the same as an AMP/CUDA crash in LwDetrImageLoss." + "reason": "Different bugs in different code paths: MoE load-balancing loss omission vs RT-DETR loss crash under fp16/CUDA." }, { "left": "issue:44623", "right": "issue:46726", "accept": false, - "reason": "save_pretrained missing files and AutoProcessor failing when a vision backend is absent are different lifecycle stages and failure causes." + "reason": "Related to processor/loading, but one is save_pretrained missing files and the other is AutoProcessor backend-missing error on load." }, { "left": "issue:43526", "right": "issue:46897", "accept": false, - "reason": "Both are preprocessing-loss style issues, but one is a Beit fast processor label reduction bug and the other is a Florence2 double-shift training-loss bug." + "reason": "Same broad label of training/processing bugs, but different models and likely different fixes; not the same underlying issue." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "A loading failure on transformers>=5.1.0 and tokenizer error masking when protobuf is missing are distinct issues." + "reason": "One is a model-loading regression in v5.1+, the other is tokenizer error masking when protobuf is absent; unrelated root causes." }, { "left": "issue:44038", "right": "issue:46506", "accept": false, - "reason": "Qwen3-VL-Moe on transformers 5.0 and PPDocLayoutV3 CPU/CUDA divergence are unrelated." + "reason": "Different subsystems and symptoms: Qwen3-VL-Moe compatibility bug vs PPDocLayoutV3 CPU/CUDA result mismatch." }, { "left": "issue:29870", "right": "issue:30333", "accept": false, - "reason": "Bart generate() UnboundLocalError and MLflow not reporting failed jobs do not share a code path." + "reason": "UnboundLocalError during generation for BART vs MLflow reporting not updated after failed training job; unrelated." }, { "left": "issue:42199", "right": "issue:46851", "accept": false, - "reason": "DETR cardinality/background-class handling is unrelated to Qwen3.5 variable-length position_ids support." + "reason": "DETR cardinality/background-class logic vs Qwen3.5 variable-length position_ids handling; different model families and code paths." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Both are warning/config-related, but one is a tied-weights warning and the other is an invalid optional post_init allowance; not the same bug." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "Memory not released on RT-DETR deletion and a MaskGenerationPipeline partial-batch bug are different pipeline behaviors." + "reason": "Memory-leak-on-delete in RT-DETR is unrelated to MaskGenerationPipeline's final partial batch handling." }, { - "left": "issue:43653", - "right": "issue:44162", + "left": "issue:44479", + "right": "issue:46170", "accept": false, - "reason": "BigBirdTokenizer special-token registration and ESM2 being broken are not the same underlying defect." + "reason": "Qwen video-input regression is unrelated to Windows/ROCm import and torch.distributed dependency problems." }, { "left": "issue:43526", "right": "issue:46894", "accept": false, - "reason": "Both mention label/loss shifting, but they are separate model implementations and do not show the same concrete bug." + "reason": "Both mention a training-loss shift pattern, but they are different model-specific reports and not enough evidence of one shared bug." }, { - "left": "issue:44060", - "right": "issue:44077", + "left": "issue:43653", + "right": "issue:44162", "accept": false, - "reason": "Incorrect tied-weights warning in Qwen3-Next and patchtsmixer post_init policy are unrelated." + "reason": "BigBirdTokenizer special-token decode bug vs ESM2 being broken; different components and failure modes." }, { - "left": "issue:44479", - "right": "issue:46170", + "left": "issue:44368", + "right": "issue:44977", "accept": false, - "reason": "A Qwen video-input regression and a Windows/ROCm distributed import error are completely different failure modes." + "reason": "Both mention Qwen3.5, but one is a tied-embeddings warning and the other is a flash-attention generation regression." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "One says NemotronHForCausalLM does not support Flash Attention 2.0 yet; the other is a generic Flash Attention 2.0 issue, with no evidence they are the same bug." - }, - { - "left": "issue:44368", - "right": "issue:44977", - "accept": false, - "reason": "A tied-word-embeddings warning and a Qwen3.5 flash-attention generation failure are different issues." + "reason": "Generic Flash Attention 2.0 topic vs NemotronHForCausalLM not supporting FA2; same feature area, different concrete issue." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are unrelated." + "reason": "Tokenizer offline cache loading and Gemma4 multi-GPU CUDA OOM are unrelated." }, { "left": "issue:44877", "right": "issue:46170", "accept": false, - "reason": "Strict config loading for granite_speech and Windows/ROCm torch.distributed import errors have no common code path." + "reason": "Config strictness for granite_speech is unrelated to Windows/ROCm import failures." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "A missing output attribute in Sam3 video encoding and a Qwen3OmniMoe video-input ValueError are different model bugs." - } - ] - }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, - "error_kind": null, - "error_message": null + "reason": "SAM 3 video attribute error vs Qwen3OmniMoe video-input ValueError; different models and code paths." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "gemma-4 ZeRO-3 from_pretrained issue is unrelated to the deepspeed optimizer-step double-call report." + }, + { + "left": "issue:46170", + "right": "issue:46832", + "accept": false, + "reason": "Windows ROCm/torch.distributed import problem vs RT-DETR forward feature-level mismatch; not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "SAM3 pixel decoder dummy Conv2d vs sam3_video/sam3_tracker config mismatch; related family, different issue." + }, + { + "left": "issue:43723", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer loading issue in v5 vs DeepSeek-R1 accuracy regression after upgrading Transformers; different symptoms and likely causes." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "Qwen3.5 tied-weights warning is unrelated to paged generate's stale num_return_sequences warning." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache get_seq_length return-type bug vs LlamaConfig head_dim validation; no shared underlying defect." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null }, { "label": "issue", @@ -13766,16 +13785,17 @@ "nodes": [ "issue:28282", "issue:30990", + "issue:41628", "issue:42994", "issue:43065", - "issue:43408", + "issue:43232", "issue:43441", "issue:43475", "issue:43482", "issue:43643", "issue:43646", + "issue:43673", "issue:43701", - "issue:43723", "issue:43906", "issue:43931", "issue:43994", @@ -13787,41 +13807,34 @@ "issue:44483", "issue:44623", "issue:44749", + "issue:44869", "issue:44871", + "issue:44933", + "issue:44964", "issue:44995", "issue:45103", + "issue:45137", "issue:45200", + "issue:45356", "issue:45375", - "issue:45397", "issue:45478", - "issue:45563", "issue:45588", "issue:45656", "issue:45685", - "issue:45987", - "issue:46082", "issue:46123", "issue:46132", - "issue:46170", "issue:46202", "issue:46421", "issue:46431", - "issue:46710", "issue:46726", "issue:46752", - "issue:46832", + "issue:46821", "issue:46851", "issue:46858" ], "soft_pairs": [ - "issue:45397|issue:45656", - "issue:43065|issue:43408", - "issue:46170|issue:46832", - "issue:43723|issue:46710", - "issue:44368|issue:45563", - "issue:45987|issue:46082", - "issue:43931|issue:44368", "issue:44995|issue:45200", + "issue:43931|issue:44368", "issue:44188|issue:46202", "issue:44483|issue:46202", "issue:43906|issue:46202", @@ -13831,8 +13844,8 @@ "issue:45103|issue:46858", "issue:43065|issue:46123", "issue:44361|issue:44749", - "issue:43994|issue:46726", "issue:44297|issue:44623", + "issue:43994|issue:46726", "issue:44749|issue:45656", "issue:43701|issue:44749", "issue:44871|issue:46132", @@ -13841,201 +13854,227 @@ "issue:43441|issue:46752", "issue:43475|issue:43643", "issue:42994|issue:43482", - "issue:28282|issue:30990", "issue:45375|issue:45478", - "issue:44188|issue:46851" + "issue:44188|issue:46851", + "issue:28282|issue:30990", + "issue:43475|issue:46821", + "issue:43232|issue:43673", + "issue:44869|issue:45356", + "issue:44964|issue:45137", + "issue:41628|issue:44933" ], - "prepared_review_unit_hash": "bdcf6ab459cc4fb1fe5e157a7eac26c5957ca8bf05f27b16dab2c60156fedd1d", + "prepared_review_unit_hash": "3d4fd37f9d29390dd5d7a3fe77dc1c5edc2f327369dfa29c6fe60e1e5e5b04f9", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36830, - "estimated_input_tokens": 9208, - "estimated_eval_tokens": 18672 + "soft_pair_count": 29, + "serialized_chars": 36459, + "estimated_input_tokens": 9115, + "estimated_eval_tokens": 18486 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "A grab bag of mostly unrelated Transformers v5 regression tickets and model-specific bugs. The soft-similarity pairs share broad themes, but none look like the same concrete defect or mergeable PR.", - "confidence": 0.73, - "canonical_issue_reason": "issue:43646 is the broadest, most representative user-facing v5 regression in the set; it captures the cluster\u2019s general theme better than the narrower model-specific reports.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers regressions. Only one soft pair looks plausibly like the same underlying bug: Qwen3.5 MoE config fields being dropped, which can surface as a from_pretrained failure. The rest are symptom-level similarities across different models/code paths and should stay separate.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45375 is the clearest root-cause report in the only plausible duplicate pair: it names the missing Qwen3_5MoeVisionConfig field directly, which is more canonical than the broader from_pretrained error in issue:45478.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44623 is a clean, concrete API regression report and a strong representative issue for this cluster of save/load failures.", + "best_issue_reason": "issue:45375 is the best anchor because it describes a concrete, fixable configuration-field omission rather than a downstream symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45397", - "right": "issue:45656", - "accept": false, - "reason": "Different failures: ZeRO/from_pretrained initialization vs duplicate optimizer stepping under deepspeed." - }, - { - "left": "issue:43065", - "right": "issue:43408", - "accept": false, - "reason": "Same SAM 3 area, but one is a dummy Conv2d implementation detail and the other is a model-type mismatch warning." - }, - { - "left": "issue:46170", - "right": "issue:46832", - "accept": false, - "reason": "Unrelated bugs: Windows/ROCm import dependency issue vs RT-DETR forward shape mismatch." - }, - { - "left": "issue:43723", - "right": "issue:46710", - "accept": false, - "reason": "Tokenizer loading failure and accuracy regression are different code-paths and different symptoms." - }, - { - "left": "issue:44368", - "right": "issue:45563", - "accept": false, - "reason": "Both are stale warning complaints, but they concern different configs and different warnings." - }, - { - "left": "issue:45987", - "right": "issue:46082", + "left": "issue:44995", + "right": "issue:45200", "accept": false, - "reason": "Cache return-type bug vs LlamaConfig validation; not the same defect." + "reason": "Different failures: stale indexer cache on second forward pass vs missing mm_token_type_ids defaulting for text-only fine-tuning." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Weight-shape mismatch on Qwen3-VL is unrelated to a tie_word_embeddings warning." - }, - { - "left": "issue:44995", - "right": "issue:45200", - "accept": false, - "reason": "Second-forward crash in GlmMoeDsa and Gemma fine-tuning token IDs are distinct issues." + "reason": "Both involve Qwen, but one is a shape mismatch on load and the other is a tie_word_embeddings warning during LoRA training; not the same bug." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "Attention-kernel divergence under torch.compile does not match a Gemma pipeline roundtrip issue." + "reason": "Different code paths: torch.compile attention-kernel divergence vs Gemma pipeline roundtrip/tokenization issue." }, { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "Chat-completions API rejection is a different bug from the Gemma pipeline roundtrip failure." + "reason": "API request rejection in chat completions is unrelated to Gemma pipeline roundtrip behavior." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "An isolated reproduction of another issue is not the same as the Gemma roundtrip bug." + "reason": "A reproduction of another issue is not the same as the Gemma pipeline roundtrip failure." }, { "left": "issue:44188", "right": "issue:46421", "accept": false, - "reason": "Compile-time attention divergence vs cache lazy initialization recompilation are different mechanisms." + "reason": "Both mention cache/compile behavior, but one is attention-kernel divergence and the other is spurious recompilation from lazy cache init; too different to merge." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "Stale indexer cache crash and flash-attention s_aux=None AttributeError are unrelated." + "reason": "Second-forward stale cache bug is unrelated to flash_attention.py crashing on s_aux=None." }, { "left": "issue:45685", "right": "issue:46431", "accept": false, - "reason": "MPS histogram support error has nothing to do with missing chat templates." + "reason": "MPS histogram/int dtype error is unrelated to missing chat template handling." }, { "left": "issue:45103", "right": "issue:46858", "accept": false, - "reason": "Docstring annotation handling and cache seq-length tensor arithmetic are separate bugs." + "reason": "Auto-docstring annotation parsing crash is a different subsystem from cache.get_seq_length() returning a device tensor in prefill." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "SAM 3 decoder construction and MaskGenerationPipeline batching are unrelated." + "reason": "Dummy Conv2d placeholder in Sam3PixelDecoder is unrelated to MaskGenerationPipeline dropping final partial batch results." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "Tokenizer AttributeError and performance slowdown after upgrade are different regressions." + "reason": "Tokenizer AttributeError and a Chinese performance regression are not the same underlying bug." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save metadata mismatch and processor save missing files are both serialization-related, but not the same concrete failure." }, { "left": "issue:43994", "right": "issue:46726", "accept": false, - "reason": "Bad results from AutoModel/pipeline and missing vision backend detection are not the same failure." + "reason": "Wrong outputs from AutoModel/pipeline are unrelated to missing vision-backend handling in AutoProcessor." }, { - "left": "issue:44297", - "right": "issue:44623", + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Performance regression during filtering is unrelated to DeepSpeed causing optimizer.step() to run twice." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch is a different problem from the reported data filtering slowdown." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id inconsistency is unrelated to AttentionInterface.register changing registered-function behavior." + }, + { + "left": "issue:43646", + "right": "issue:44162", "accept": false, - "reason": "Related save_pretrained area, but tokenizer_class metadata mismatch is distinct from missing processor files." + "reason": "Both are v5 breakages, but one is custom model initialization and the other is ESM2 breakage; too broad to treat as the same bug." }, { "left": "issue:44623", "right": "issue:45656", "accept": false, - "reason": "Processor save_pretrained file loss and double optimizer stepping are unrelated." + "reason": "Processor save_pretrained missing files is unrelated to duplicated optimizer steps under DeepSpeed." }, { "left": "issue:43441", "right": "issue:46752", "accept": false, - "reason": "FlashAttention crash on Ministral-3 and empty-list chat template IndexError are different code paths." + "reason": "FlashAttention failure for Ministral-3 is unrelated to apply_chat_template crashing on an empty list." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "Missing SAM 3 vision output field and trust_remote_code config field loss are not the same bug." + "reason": "Missing fpn_position_embeddings in SAM 3 video is not the same as trust_remote_code causing missing fields in AutoConfig output." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized saving failure and GGUF loading failure are separate model I/O problems." + "reason": "Quantized model saving failure and GGUF loading failure are opposite directions and different code paths." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": true, + "reason": "These appear to describe the same Qwen3.5 MoE config bug: a missing deepstack_visual_indexes field being dropped by strict validation, which then surfaces as a from_pretrained error in newer transformers." + }, + { + "left": "issue:44188", + "right": "issue:46851", + "accept": false, + "reason": "Attention-kernel branching under torch.compile is unrelated to Qwen3.5 variable-length position_ids handling." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "Missing PyTorch import error and Sentence Transformers hanging on load are different issues." + "reason": "PyTorch import error for AutoModel is unrelated to Sentence Transformers getting stuck while loading." }, { - "left": "issue:45375", - "right": "issue:45478", + "left": "issue:43475", + "right": "issue:46821", "accept": false, - "reason": "Missing strict config field and a v5.4.0 from_pretrained failure on Qwen3.5 Moe are related by model family only, not the same defect." + "reason": "SAM 3 video missing an attribute is unrelated to TimesFM window_size raising AttributeError." }, { - "left": "issue:44188", - "right": "issue:46851", + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both touch generation internals, but sync_gpus/update_model_kwargs and chunked_prefill cache initialization are distinct bugs." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decode crash and Kimi tokenizer codec regression are different tokenizer/audio issues." + }, + { + "left": "issue:44964", + "right": "issue:45137", "accept": false, - "reason": "torch.compile attention-kernel divergence is unrelated to variable-length position_ids handling in Qwen3.5." + "reason": "Phi-4 multimodal loading failure is unrelated to the DeepSpeed ZeRO3 deque IndexError." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Missing AutoImageProcessor import and a nonexistent image_utils import are similar in theme but not the same bug." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet: it stays at a high level, distinguishes the one plausible Qwen3.5 MoE config duplicate from the unrelated regressions, and the soft-edge verdicts are generally conservative. The accepted 45375/45478 pair is reasonably supported by the titles and is phrased cautiously enough." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -14055,41 +14094,42 @@ "issue:36010", "issue:39692", "issue:41628", + "issue:42175", "issue:42371", "issue:42673", + "issue:42831", "issue:42994", "issue:43232", "issue:43278", - "issue:43475", + "issue:43504", "issue:43525", - "issue:43673", "issue:43704", "issue:43723", "issue:43784", + "issue:43824", "issue:43873", "issue:43986", "issue:44038", - "issue:44188", "issue:44222", "issue:44351", "issue:44373", "issue:44623", + "issue:44792", "issue:44857", - "issue:44869", "issue:44933", "issue:44945", - "issue:44964", "issue:45081", "issue:45092", "issue:45103", "issue:45137", "issue:45230", "issue:45310", - "issue:45356", "issue:45375", "issue:45397", "issue:45406", + "issue:45561", "issue:45593", + "issue:45600", "issue:45753", "issue:46202", "issue:46326", @@ -14097,27 +14137,20 @@ "issue:46463", "issue:46701", "issue:46726", - "issue:46821", "issue:46851", "issue:46877" ], "soft_pairs": [ - "issue:43475|issue:46821", - "issue:43232|issue:43673", - "issue:44964|issue:45137", - "issue:44869|issue:45356", - "issue:39692|issue:46726", - "issue:41628|issue:44933", "issue:45310|issue:45375", "issue:46463|issue:46877", - "issue:43986|issue:46726", + "issue:39692|issue:46726", "issue:44373|issue:46851", "issue:45230|issue:45397", + "issue:43986|issue:46726", "issue:44038|issue:44623", "issue:43723|issue:46326", "issue:42994|issue:43525", "issue:42371|issue:45092", - "issue:43232|issue:44188", "issue:43873|issue:44945", "issue:45406|issue:46202", "issue:43784|issue:46399", @@ -14126,119 +14159,95 @@ "issue:45593|issue:46877", "issue:41628|issue:44351", "issue:36010|issue:41628", - "issue:46202|issue:46701", "issue:45081|issue:45103", + "issue:46202|issue:46701", "issue:43278|issue:45137", "issue:44038|issue:45310", "issue:42673|issue:43704", - "issue:44222|issue:45310" + "issue:44222|issue:45310", + "issue:42994|issue:43824", + "issue:42831|issue:44222", + "issue:43504|issue:45561", + "issue:43232|issue:44792", + "issue:45561|issue:45600", + "issue:42175|issue:44933" ], - "prepared_review_unit_hash": "0c85474418513c5ce5392c069bbcd55e32e03b6cb1d34b9716bb1ffc20d2f5e4", + "prepared_review_unit_hash": "aad4afa711e752f46473039376d29b0b50d7ffdbdf799bca00b8bc0ae9d3e70c", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36406, - "estimated_input_tokens": 9102, - "estimated_eval_tokens": 18460 + "serialized_chars": 36309, + "estimated_input_tokens": 9078, + "estimated_eval_tokens": 18412 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly heterogeneous; only the Qwen3ForCausalLM VRAM leak reports are a clear duplicate. Most other soft matches are broad similarity matches around imports, tokenizers, or generation APIs but describe different concrete failures and should not be merged.", - "confidence": 0.87, - "canonical_issue_reason": "issue:42673 is the clearest canonical report: it is a concrete, well-scoped VRAM leak bug and is an exact duplicate of issue:43704 by title.", + "summary": "This cluster is mostly a heterogeneous set of Transformers bug reports spanning unrelated subsystems. The only clear duplicate pair is the Qwen3ForCausalLM VRAM leak report (42673/43704); the rest are too loosely related to merge safely.", + "confidence": 0.93, + "canonical_issue_reason": "issue:42673 is the more complete report for the exact duplicate VRAM leak/threading bug and has substantially more discussion than issue:43704.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42673 is the strongest issue to anchor the cluster because it is the only exact duplicate pair here and has the fuller discussion/activity footprint.", + "best_issue_reason": "issue:42673 is the best canonical issue: it precisely describes the bug, has richer context, and cleanly subsumes issue:43704.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43475", - "right": "issue:46821", - "accept": false, - "reason": "Different models and different AttributeErrors; no evidence they share the same code path." - }, - { - "left": "issue:43232", - "right": "issue:43673", - "accept": false, - "reason": "Both involve generation internals, but one is about sync_gpus/_update_model_kwargs and the other about chunked_prefill cache handling; not the same bug." - }, - { - "left": "issue:44964", - "right": "issue:45137", - "accept": false, - "reason": "Unrelated failures: model loading for Phi-4-multimodal versus a DeepSpeed ZeRO3 deque error." - }, - { - "left": "issue:44869", - "right": "issue:45356", - "accept": false, - "reason": "Both are tokenizer-related, but they affect different models and different failure modes; not duplicate evidence." - }, - { - "left": "issue:39692", - "right": "issue:46726", - "accept": false, - "reason": "One is a SigLIP2 doc example/model-processor mismatch; the other is an AutoProcessor backend-missing error." - }, - { - "left": "issue:41628", - "right": "issue:44933", - "accept": false, - "reason": "Both are import problems, but for different symbols/modules; too broad to be the same underlying bug." - }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Likely related Qwen3.5/Moe loading issues, but one is a generic from_pretrained error and the other is a specific missing config field; not enough to treat as the same bug." + "reason": "Same model family, but one is a from_pretrained error and the other is a missing config field; not clearly the same concrete bug." }, { "left": "issue:46463", "right": "issue:46877", "accept": false, - "reason": "A decoder-input validation error and an auto_docstring import-time logging issue are unrelated." + "reason": "Unrelated problems: T5 decoder-input validation messaging vs auto_docstring stdout diagnostics." }, { - "left": "issue:43986", + "left": "issue:39692", "right": "issue:46726", "accept": false, - "reason": "Video AutoProcessor missing torchvision handling is unrelated to the SigLIP2 documentation/quantization issue." + "reason": "Both involve processor/vision docs or backend handling, but they are different failure modes and code paths." }, { "left": "issue:44373", "right": "issue:46851", "accept": false, - "reason": "Docstring mismatch versus a Qwen3.5 variable-length input runtime bug; different problem domains." + "reason": "Both mention position_ids, but one is a docstring issue and the other is a runtime bug in Qwen3.5 variable-length handling." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report versus a specific gemma-4 ZeRO3 loading issue; no concrete overlap." + "reason": "Generic placeholder bug report versus a specific Gemma-4 ZeRO3 regression; no evidence they are the same issue." + }, + { + "left": "issue:43986", + "right": "issue:46726", + "accept": false, + "reason": "Both concern missing vision backends, but one is a video-model AutoProcessor crash and the other is an unrecognized image-processor error." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe compatibility bug and processor.save_pretrained missing files are different issues." + "reason": "Different bugs: Qwen3-VL-Moe loading versus processor.save_pretrained missing files." }, { "left": "issue:43723", "right": "issue:46326", "accept": false, - "reason": "Tokenizer loading in v5 and continue_final_message behavior with mistral_common are distinct failures." + "reason": "Tokenizer loading failure in v5 and a mistral_common continue_final_message bug are separate problems." }, { "left": "issue:42994", @@ -14250,103 +14259,133 @@ "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 API guidance and InternVL2 meta-initialization compatibility are not the same bug." - }, - { - "left": "issue:43232", - "right": "issue:44188", - "accept": false, - "reason": "Generation cache handling and attention-kernel divergence under torch.compile are separate code paths." + "reason": "TF32 API guidance and InternVL2 meta-initialization incompatibility are unrelated." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading behavior and pipeline-parallel wrong outputs are different issues." + "reason": "Quantization offloading behavior and pipeline-parallel output correctness are different code paths." }, { "left": "issue:45406", "right": "issue:46202", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and Gemma pipeline roundtrip on \u2581 are not the same defect." + "reason": "Gemma processor attribute crash and Gemma pipeline roundtrip failure are distinct bugs." }, { "left": "issue:43784", "right": "issue:46399", "accept": false, - "reason": "Sentence-transformers import NameError and Qwen3.5 kernel registration crash are unrelated." + "reason": "Import-time NameError in sentence-transformers integration versus a Qwen3.5 kernel registration crash." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Pipeline-parallel output corruption and a Qwen3.5 infinite-loop image bug are different failures." + "reason": "Pipeline-parallel output corruption and a Qwen3_5 image-specific infinite loop do not match." }, { "left": "issue:44857", "right": "issue:46877", "accept": false, - "reason": "AMP/CUDA crash in LwDetrImageLoss versus auto_docstring stdout diagnostics are unrelated." + "reason": "CUDA AMP loss crash and auto_docstring import-time logging are unrelated." }, { "left": "issue:45593", "right": "issue:46877", "accept": false, - "reason": "Missing auxiliary losses in D-FINE and auto_docstring import-time logging are unrelated." + "reason": "D-FINE auxiliary-loss behavior and docstring diagnostics are unrelated." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they concern different public APIs and likely different fixes." + "reason": "Both are import errors, but they are for different symbols with different likely fixes." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "GenerationMixin import failure and AutoImageProcessor import failure are separate public API regressions." + "reason": "Different missing exports/imports from different parts of transformers; not the same bug." }, { - "left": "issue:46202", - "right": "issue:46701", + "left": "issue:45081", + "right": "issue:45103", "accept": false, - "reason": "Gemma pipeline roundtrip bug and DiffusionGemmaGenerationOutput pipeline failure are related only by model family, not by exact bug." + "reason": "Two separate helper-function crashes with different triggers and stack paths." }, { - "left": "issue:45081", - "right": "issue:45103", + "left": "issue:46202", + "right": "issue:46701", "accept": false, - "reason": "Mistral tokenizer regex crash and auto_docstring future-annotations crash are unrelated AttributeErrors." + "reason": "Both mention Gemma, but one is a tokenizer roundtrip issue and the other is an image-text-to-text pipeline failure." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "Dtype drift in evaluation and a ZeRO3 deque IndexError are unrelated." + "reason": "Embedding dtype drift in eval and a DeepSpeed ZeRO3 deque IndexError are unrelated." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Different Qwen MoE variants and different symptoms; too broad to be the same concrete issue." + "reason": "Related model family, but the reported failures differ: one is a Qwen3-VL-Moe bug and the other is a Qwen3.5 MoE from_pretrained regression." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Exact same title and same VRAM leak symptom for Qwen3ForCausalLM in multiple dataloader threads; clear duplicate." + "reason": "Exact same VRAM leak report for Qwen3ForCausalLM used across multiple dataloader threads; 43704 is a duplicate of 42673." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "FP8 save_pretrained bug and Qwen3.5 MoE from_pretrained error are different failure modes." - } - ] + "reason": "FP8 save_pretrained failure and Qwen3.5 Moe from_pretrained loading error are different concrete problems." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized save failure and missing Qwen2_5_VL export are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both are FP8-related, but one is an accuracy regression and the other is a save_pretrained bug." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Beit legacy-field loading bug and pytest-xdist race on captured_info.txt are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and janus image generation test failure are different issues." + }, + { + "left": "issue:45561", + "right": "issue:45600", + "accept": false, + "reason": "Testing-utils xdist race and broken auto_mappings references are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Backend packaging omission for TensorFlow and a missing image_utils import are unrelated." + } + ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded and conservative overall. The canonical choice is supported by the exact title duplicate between issue:42673 and issue:43704, and the rest of the soft-edge rejections stay appropriately cautious given only loose topical similarity in the packet." + "feedback": "The canonical choice and cluster summary are grounded in the packet: 42673 and 43704 have identical titles and the discussion counts support choosing 42673 as canonical. The soft-edge verdicts are also conservative overall, with no strong overclaims beyond what the titles support." }, "evaluator_used": true, "retried": false, @@ -14371,17 +14410,16 @@ "issue:30333", "issue:30990", "issue:33290", + "issue:34689", "issue:35141", "issue:39692", "issue:42175", + "issue:42222", "issue:42673", - "issue:42831", "issue:42898", - "issue:42994", - "issue:43232", + "issue:43116", "issue:43295", "issue:43479", - "issue:43504", "issue:43723", "issue:43784", "issue:43824", @@ -14389,22 +14427,24 @@ "issue:43976", "issue:44016", "issue:44079", - "issue:44222", + "issue:44190", + "issue:44295", "issue:44496", "issue:44568", "issue:44623", "issue:44625", "issue:44734", "issue:44779", - "issue:44792", "issue:44869", "issue:44933", "issue:45092", "issue:45103", + "issue:45200", "issue:45310", + "issue:45356", "issue:45405", "issue:45561", - "issue:45600", + "issue:45656", "issue:45712", "issue:45812", "issue:46077", @@ -14416,12 +14456,6 @@ "issue:46909" ], "soft_pairs": [ - "issue:42994|issue:43824", - "issue:42831|issue:44222", - "issue:43504|issue:45561", - "issue:45561|issue:45600", - "issue:43232|issue:44792", - "issue:42175|issue:44933", "issue:44496|issue:45561", "issue:43976|issue:44933", "issue:45405|issue:46877", @@ -14433,8 +14467,8 @@ "issue:44623|issue:45310", "issue:46077|issue:46123", "issue:42673|issue:46693", - "issue:42898|issue:45812", "issue:43479|issue:44869", + "issue:42898|issue:45812", "issue:43723|issue:44568", "issue:44016|issue:46231", "issue:44625|issue:46877", @@ -14445,216 +14479,222 @@ "issue:44734|issue:45103", "issue:43874|issue:45092", "issue:27957|issue:30333", - "issue:42898|issue:43295" + "issue:42898|issue:43295", + "issue:42222|issue:45200", + "issue:30333|issue:34689", + "issue:42898|issue:45356", + "issue:44295|issue:45656", + "issue:43116|issue:44190", + "issue:45405|issue:45712" ], - "prepared_review_unit_hash": "fe9d2f6a32e86493f00c96374aef8e26a13b0c13aeea3bd0f7be527e8dbab173", + "prepared_review_unit_hash": "a98a17444645501614cb7842ef1b576711da33c96c324652719f3afe1e003849", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36211, - "estimated_input_tokens": 9053, - "estimated_eval_tokens": 18362 + "serialized_chars": 36964, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: tokenizer regressions, multimodal/model-loading bugs, training/runtime issues, docs problems, and packaging/import failures. None of the soft pairs look like the same underlying bug or change, so no duplicate merge is warranted.", - "confidence": 0.95, - "canonical_issue_reason": null, + "summary": "These items are mostly unrelated Transformers issues with only superficial similarity around regressions, tokenization, multimodal loading, and tooling. I do not see any soft pair that looks like the same underlying bug/change.", + "confidence": 0.88, + "canonical_issue_reason": "If a single anchor is needed, issue 45812 is the strongest representative: it is a concrete, high-severity silent regression affecting all Granite tokenizers.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 45812 is the most actionable and broadly impactful report in the set, with a clear user-facing correctness bug and reproducible symptoms.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42994", - "right": "issue:43824", - "accept": false, - "reason": "Quantized model saving failure vs missing Qwen2_5_VLForConditionalGeneration import; unrelated bugs." - }, - { - "left": "issue:42831", - "right": "issue:44222", - "accept": false, - "reason": "Accuracy regression in FineGrainedFP8 vs FP8 save_pretrained failure for MoE; different failure modes." - }, - { - "left": "issue:43504", - "right": "issue:45561", - "accept": false, - "reason": "Beit pretrained preset loading bug vs pytest-xdist file race; unrelated subsystems." - }, - { - "left": "issue:45561", - "right": "issue:45600", - "accept": false, - "reason": "Captured_info.txt race in tests vs stale auto_mappings config references; different code paths." - }, - { - "left": "issue:43232", - "right": "issue:44792", - "accept": false, - "reason": "Generation kwargs update bug vs janus image generation test failure; not the same issue." - }, - { - "left": "issue:42175", - "right": "issue:44933", - "accept": false, - "reason": "Missing TensorFlow backend in install extra vs nonexistent image_utils import; unrelated." - }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Unrecognized model config error vs test file race; no overlap beyond being closed issues." + "reason": "Different subsystems and failure modes: model loading/config resolution vs pytest-xdist file-race behavior." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python version compatibility issue vs image_utils import problem; unrelated." + "reason": "Unrelated problems: Python compatibility claim vs a bad import path in image_utils." }, { "left": "issue:45405", "right": "issue:46877", "accept": false, - "reason": "PEFT minimum version bump vs auto_docstring stdout diagnostics; different maintenance issues." + "reason": "Release/versioning complaint vs import-time docstring lint output; no shared bug." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLflow reporting on failed jobs vs OOM with Adafactor/Deepspeed; both training-related but distinct bugs." + "reason": "MLflow reporting after failure and DeepSpeed optimizer OOM are distinct training issues." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "SentenceTransformers loading stall vs embedding resize/reinit bug; unrelated." + "reason": "Loading hang in sentence-transformers does not match embedding reinitialization in post_init." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Different tokenizer regressions: cleanup-space behavior vs missing BOS/EOS on add_special_tokens." + "reason": "Both are tokenizer-related, but one is v5 cleanup-space behavior and the other is special-token insertion for a specific tokenizer." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "Packaging/backend dependency issue vs leftover dummy classes in repo checks; unrelated." + "reason": "Missing TensorFlow in an install extra is unrelated to leftover dummy PT classes." }, { "left": "issue:43479", "right": "issue:46829", "accept": false, - "reason": "Phi4 multimodal config default initialization vs VideoPrism output field mix-up; different model families and bugs." + "reason": "Different model families and bugs: default config initialization vs wrong output field assignment." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Missing processor.save_pretrained files vs Qwen3.5 MoE from_pretrained error; not the same defect." + "reason": "Saving processor artifacts missing files is not the same as a Qwen3.5 MoE from_pretrained error." }, { "left": "issue:46077", "right": "issue:46123", "accept": false, - "reason": "Attention-mask construction bug vs partial-batch pipeline bookkeeping bug; unrelated." + "reason": "Attention-mask construction for Gemma4 is unrelated to final-partial-batch handling in MaskGenerationPipeline." }, { "left": "issue:42673", "right": "issue:46693", "accept": false, - "reason": "VRAM leak with multiple dataloader threads vs FA2 long-context performance regression; different symptoms and causes." + "reason": "A VRAM leak in multithreaded dataloading is not the same as a long-context FA2 performance regression." }, { - "left": "issue:42898", - "right": "issue:45812", + "left": "issue:43479", + "right": "issue:44869", "accept": false, - "reason": "Both are tokenizer regressions, but one is cleanup spacing and the other is wrong Granite token IDs; distinct bugs." + "reason": "Phi4 multimodal config defaults and Whisper timestamp decode crash are different code paths." }, { - "left": "issue:43479", - "right": "issue:44869", + "left": "issue:42898", + "right": "issue:45812", "accept": false, - "reason": "Multimodal config default handling vs Whisper word timestamp decode crash; unrelated." + "reason": "Both affect tokenization behavior in v5, but one is cleanup-space semantics and the other is wrong token IDs for Granite models." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "AutoTokenizer v5 loading issue vs add_special_tokens BOS/EOS regression; same area, different concrete behavior." + "reason": "Tokenizer loading in v5 and add_special_tokens behavior for mdeberta-v3 are separate regressions." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error vs docs parameter name mismatch; unrelated documentation problems." + "reason": "Notebook syntax error and documentation parameter-name mismatch are unrelated doc issues." }, { "left": "issue:44625", "right": "issue:46877", "accept": false, - "reason": "Qwen3.5 num_labels propagation bug vs docstring lint stdout noise; different areas." + "reason": "num_labels propagation in Qwen3.5 has nothing to do with auto_docstring logging diagnostics." }, { "left": "issue:39692", "right": "issue:46909", "accept": false, - "reason": "SigLIP2 example with model/processor mismatch vs DIA doc example broken; both docs, but not the same example or bug." + "reason": "Two different broken documentation examples with different models and failures." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Two tokenizer regressions, but different models and symptoms; not mergeable as one fix." + "reason": "Both are tokenizer regressions, but the affected models and failure mechanisms differ." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Multimodal config defaults vs ModelOutput key assignment bug; unrelated." + "reason": "Config defaulting in Phi4Multimodal and ModelOutput key assignment are unrelated internals." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Import error in sentence-transformers due to missing nn vs missing Qwen2_5_VLForConditionalGeneration import; unrelated imports." + "reason": "Import-time NameError in sentence-transformers is not the same as a missing Qwen2.5-VL export." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Serve KV-cache indexing crash vs auto_docstring future-annotations crash; different code paths." + "reason": "Serve KV-cache indexing crash and auto_docstring future-annotations crash are separate code paths." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Missing image-patch helper in GLM46V processor vs InternVL2 meta-init incompatibility; different multimodal loading bugs." + "reason": "Missing image-patch method in GLM46V and InternVL2 meta-init incompatibility are different multimodal bugs." }, { "left": "issue:27957", "right": "issue:30333", "accept": false, - "reason": "FA2 support for XLMRoberta vs MLflow reporting after failed training; unrelated." + "reason": "Flash Attention support for XLM-Roberta and MLflow reporting after training failure are unrelated." }, { "left": "issue:42898", "right": "issue:43295", "accept": false, - "reason": "Tokenizer cleanup regression vs processor.tokenizer/image handling regression; different API breakages." + "reason": "Tokenization-space behavior and processor/tokenizer/image regression in a custom model are not the same bug." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Broken VitPose model files and Gemma4 mm_token_type_ids defaults are separate multimodal issues." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLflow failure reporting and Llama 3.2 Vision model-loading break are unrelated regressions." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Different tokenizer regressions affecting different models and codecs." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position_ids buffer access error and duplicate optimizer-step behavior with DeepSpeed are distinct problems." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Empty multi-label classification results and local dataset loading failure are different example-script issues." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "Unreleased dependency pin and leftover dummy class cleanup are unrelated maintenance issues." } ] }, @@ -14681,16 +14721,16 @@ "issue:22355", "issue:28282", "issue:30333", - "issue:34689", + "issue:30990", "issue:36331", + "issue:38175", "issue:41628", - "issue:42222", "issue:42757", "issue:42796", "issue:42831", - "issue:42898", - "issue:43116", "issue:43122", + "issue:43232", + "issue:43475", "issue:43479", "issue:43644", "issue:43723", @@ -14698,7 +14738,7 @@ "issue:44075", "issue:44190", "issue:44222", - "issue:44295", + "issue:44263", "issue:44297", "issue:44360", "issue:44373", @@ -14709,30 +14749,25 @@ "issue:44933", "issue:44977", "issue:45071", + "issue:45092", "issue:45125", - "issue:45200", "issue:45245", "issue:45310", - "issue:45356", - "issue:45405", "issue:45656", "issue:45712", + "issue:45834", "issue:46082", "issue:46132", "issue:46209", "issue:46231", + "issue:46305", + "issue:46424", "issue:46614", "issue:46710", "issue:46762", "issue:46932" ], "soft_pairs": [ - "issue:42222|issue:45200", - "issue:30333|issue:34689", - "issue:42898|issue:45356", - "issue:44295|issue:45656", - "issue:43116|issue:44190", - "issue:45405|issue:45712", "issue:30333|issue:36331", "issue:22355|issue:28282", "issue:44222|issue:44297", @@ -14745,8 +14780,8 @@ "issue:42831|issue:44623", "issue:44297|issue:45656", "issue:45245|issue:46132", - "issue:44222|issue:44623", "issue:43479|issue:43644", + "issue:44222|issue:44623", "issue:46132|issue:46710", "issue:44373|issue:44485", "issue:41628|issue:42757", @@ -14756,138 +14791,107 @@ "issue:46614|issue:46932", "issue:44222|issue:44484", "issue:43122|issue:46209", - "issue:44190|issue:46082" + "issue:44190|issue:46082", + "issue:44263|issue:45092", + "issue:43232|issue:46424", + "issue:43232|issue:45834", + "issue:43475|issue:46305", + "issue:30990|issue:38175" ], - "prepared_review_unit_hash": "8b732b145c35d938a68f826195219ac2eaeab777e7ce84f2ef40ab78c27897e1", + "prepared_review_unit_hash": "faa3ed36ba6d8733ba2a09f561681f046d9d5ff142bf502711e3750649e06cb1", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 35832, - "estimated_input_tokens": 8958, - "estimated_eval_tokens": 18172 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36037, + "estimated_input_tokens": 9010, + "estimated_eval_tokens": 18276 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is a loose semantic cluster of unrelated Transformers issues; most are model-, tokenizer-, config-, or tooling-specific regressions that only share generic language around imports, loading, or v5 behavior. No soft pair looks like the same underlying bug/change, so none should be merged as duplicates.", - "confidence": 0.91, - "canonical_issue_reason": "Issue 43723 is the closest thing to a cluster center because it is a broad core-loading regression report ('AutoTokenizer.from_pretrained in v5') rather than a very narrow model-specific or documentation-only bug. That said, the cluster does not contain a true duplicate set.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues; the soft pairs are only superficially similar and do not describe the same underlying bug/change. No PRs are present.", + "confidence": 0.43, + "canonical_issue_reason": "No single issue is a clear canonical duplicate target because the set spans many unrelated regressions, import errors, tokenizer/model bugs, and version-migration problems.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43723 is the most generally reusable representative among these reports; it targets a common API path and is easier to triage as a canonical reference than the narrower, more specialized issues.", + "best_issue_reason": "issue:45071 is the best representative only in the loose sense that it is a broad Transformers v5 regression report; however, it is not a true duplicate of the other issues.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42222", - "right": "issue:45200", - "accept": false, - "reason": "Different models and different failure modes: vitpose model breakage vs Gemma 4 mm_token_type_ids defaults." - }, - { - "left": "issue:30333", - "right": "issue:34689", - "accept": false, - "reason": "Unrelated subsystems: MLflow job-finalization reporting vs Llama 3.2 vision model loading regression." - }, - { - "left": "issue:42898", - "right": "issue:45356", - "accept": false, - "reason": "Both are tokenizer regressions, but the concrete bugs differ: global clean_up_tokenization_spaces behavior vs Kimi-K2.5 codec/warning handling." - }, - { - "left": "issue:44295", - "right": "issue:45656", - "accept": false, - "reason": "No shared code-path problem: position_ids buffer access vs optimizer stepping twice under deepspeed." - }, - { - "left": "issue:43116", - "right": "issue:44190", - "accept": false, - "reason": "Both involve example scripts, but they are different scripts and different bugs (multi-label classification output vs local dataset loading)." - }, - { - "left": "issue:45405", - "right": "issue:45712", - "accept": false, - "reason": "Release/version requirement mismatch is unrelated to dummy class cleanup in dummy_pt_objects.py." - }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "Different training-stack issues: MLflow reporting on failure vs CustomTrainer compute_loss signature change." + "reason": "Different failures: MLflow run-state reporting after training job failure vs. a CustomTrainer method signature error." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Both are import errors, but one is a missing transformers.onnx module and the other is a missing PyTorch dependency." + "reason": "Both are import errors, but for different missing modules/libraries and different root causes." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "Both mention save_pretrained, but one is FP8 MoE saving and the other is tokenizer config metadata mismatch." + "reason": "Both touch FP8/tokenizer save flows, but one is about MOE FP8 save_pretrained while the other is a tokenizer_config class mismatch." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "Optimizer behavior bug and documentation typo are not the same underlying change." + "reason": "Optimizer arguments not being used is unrelated to a docs typo about set_training parameter naming." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Same general model family, but different concrete bugs: tied-weights warning vs missing tensor-parallel plan." + "reason": "Both are Qwen3-family MoE issues, but one is an incorrect tied-weights warning and the other is missing tensor-parallel planning." }, { "left": "issue:44448", "right": "issue:46132", "accept": false, - "reason": "Different issues in different code paths: Pegasus output regression vs AttentionInterface registration semantics." + "reason": "Different problems: output regression across versions vs. AttentionInterface.register changing registered-function behavior." }, { "left": "issue:44933", "right": "issue:45712", "accept": false, - "reason": "Both touch imports/dummy objects, but one is a missing import from image_utils and the other is stale dummy class leakage." + "reason": "Both concern missing/nonexistent imports, but they affect different modules and different cleanup issues." }, { "left": "issue:42796", "right": "issue:43723", "accept": false, - "reason": "Both are tokenizer-related, but one is BPE-marker handling and the other is loading AutoTokenizer in v5; not the same bug." + "reason": "Tokenizer BPE-marker handling and AutoTokenizer loading failures are distinct bugs." }, { "left": "issue:44360", "right": "issue:46762", "accept": false, - "reason": "Numerical discrepancy reports, but they concern different attention mechanisms and different root causes." + "reason": "Both mention numerical behavior in attention-like code, but they are different implementations with different observed problems." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "Accuracy loss from FineGrainedFP8 is unrelated to missing files in processor.save_pretrained." + "reason": "FineGrainedFP8 accuracy and processor.save_pretrained missing files are unrelated." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer save metadata mismatch and double optimizer steps are unrelated." + "reason": "Tokenizer save-pretrained metadata mismatch is unrelated to optimizer step being called twice under deepspeed." }, { "left": "issue:45245", @@ -14896,76 +14900,106 @@ "reason": "Category-count runtime error is unrelated to AttentionInterface.register behavior changes." }, { - "left": "issue:44222", - "right": "issue:44484", + "left": "issue:43479", + "right": "issue:43644", "accept": false, - "reason": "Both involve saving models, but one is FP8 MoE save_pretrained behavior and the other is a question about shard size defaults." + "reason": "Config default initialization bug and non-persistent buffer corruption are different code paths." }, { - "left": "issue:43479", - "right": "issue:43644", + "left": "issue:44222", + "right": "issue:44623", "accept": false, - "reason": "Config initialization with None and non-persistent buffer junk are different bugs in different mechanisms." + "reason": "Both involve saving artifacts, but one is FP8 MOE model persistence and the other is processor save file completeness." }, { "left": "issue:46132", "right": "issue:46710", "accept": false, - "reason": "An API behavior change in AttentionInterface is too broad to be the same as a specific model accuracy regression." + "reason": "AttentionInterface.register semantics and an accuracy regression after upgrade are not the same bug." }, { "left": "issue:44373", "right": "issue:44485", "accept": false, - "reason": "Docstring correction and GLM-5 RoPE implementation are not the same code-path problem." + "reason": "Docstring mismatch for position_ids is unrelated to GLM-5 RoPE implementation behavior." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Different missing imports from different packages: transformers vs huggingface_hub." + "reason": "Both are import-related, but they concern different symbols from different packages and likely different breakages." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer save-pretrained metadata mismatch is not the same bug as a Qwen3.5 MoE from_pretrained failure." + "reason": "Tokenizer save-pretrained metadata mismatch is not the same as Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:45071", "right": "issue:46082", "accept": false, - "reason": "Both are config-related regressions, but they are distinct: PretrainedConfig type checking vs LlamaConfig head_dim validation." + "reason": "PretrainedConfig type-checking regression and LlamaConfig head_dim validation are different configuration bugs." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Tokenizer loading regression and Qwen3.5 flash-attention generation issue are different failure modes." + "reason": "AutoTokenizer loading failure and Qwen3.5 flash-attention generation issues are unrelated." }, { "left": "issue:46614", "right": "issue:46932", "accept": false, - "reason": "Phonemizer backend state bug is unrelated to Gemma4 quantized embedder dtype casting." + "reason": "Both are tokenizer/processing-related multilingual issues, but one is stale phonemizer state and the other is Gemma4 quantized embedder dtype casting." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "Repeated pair; still not the same underlying bug: FP8 MoE save_pretrained vs shard-size default question." + "reason": "FP8 save_pretrained moe and max_shard_size default sizing are different save-path concerns with different symptoms." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenization regression across versions is unrelated to FP8 inference failure on B200." + "reason": "Tokenization differences across versions and FP8 inference failure on B200 are unrelated." }, { "left": "issue:44190", "right": "issue:46082", "accept": false, - "reason": "Example-script dataset loading bug and LlamaConfig validation bug are unrelated." + "reason": "Local dataset loading in a training script is unrelated to LlamaConfig head_dim validation." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return values in a specific indexer are unrelated to remote-code checkpoint incompatibility with meta initialization." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and cache max length behavior are separate generation/cache bugs." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs update bug is unrelated to Kosmos2.5 OCR index error." + }, + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "SAM 3 Video missing attribute and MoonshineStreaming off-by-one mask logic are distinct model bugs." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang and Siglip2 zero-probabilities are different model-specific issues." } ] }, @@ -14993,17 +15027,18 @@ "issue:27957", "issue:28218", "issue:30990", - "issue:38175", "issue:39401", "issue:41628", + "issue:42757", + "issue:42796", "issue:42831", "issue:42886", "issue:42907", "issue:42994", "issue:43066", + "issue:43122", "issue:43232", "issue:43408", - "issue:43475", "issue:43644", "issue:43824", "issue:44038", @@ -15019,12 +15054,13 @@ "issue:44623", "issue:44829", "issue:45092", + "issue:45278", "issue:45310", "issue:45357", "issue:45381", "issue:45406", + "issue:45593", "issue:45753", - "issue:45834", "issue:46002", "issue:46050", "issue:46209", @@ -15032,33 +15068,27 @@ "issue:46328", "issue:46369", "issue:46402", - "issue:46424", "issue:46431", "issue:46655", "issue:46909" ], "soft_pairs": [ - "issue:44263|issue:45092", - "issue:43232|issue:46424", - "issue:43232|issue:45834", - "issue:43475|issue:46305", - "issue:30990|issue:38175", "issue:46305|issue:46655", "issue:46209|issue:46431", "issue:44295|issue:44623", "issue:44263|issue:44360", - "issue:45092|issue:45357", "issue:44263|issue:44373", + "issue:45092|issue:45357", "issue:42907|issue:46050", "issue:46369|issue:46909", "issue:42831|issue:45310", "issue:45381|issue:46305", - "issue:44483|issue:45406", "issue:44279|issue:46328", + "issue:44483|issue:45406", "issue:22355|issue:27957", - "issue:44829|issue:46002", "issue:43644|issue:46305", "issue:44261|issue:45092", + "issue:44829|issue:46002", "issue:42831|issue:44038", "issue:43408|issue:46402", "issue:41628|issue:43824", @@ -15068,222 +15098,220 @@ "issue:42886|issue:43066", "issue:27957|issue:30990", "issue:42886|issue:42994", - "issue:44188|issue:44373" + "issue:44188|issue:44373", + "issue:42757|issue:45278", + "issue:43408|issue:46369", + "issue:43232|issue:45593", + "issue:42796|issue:43122" ], - "prepared_review_unit_hash": "575e796a5ce885d4f52d7a68e154c1651e7f6e54943ef425f6cd43ec26a7664b", + "prepared_review_unit_hash": "5928d1736752922c95533b2845d253d98e167b72528081179d72b44620613e94", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 35956, - "estimated_input_tokens": 8989, - "estimated_eval_tokens": 18234 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36190, + "estimated_input_tokens": 9048, + "estimated_eval_tokens": 18352 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes unrelated tokenizer, model-loading, quantization, cache, attention, and documentation issues across different models and code paths. None of the soft-linked pairs look like the same underlying bug, so all are rejected.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue fits this cluster: the items are unrelated enough that the soft links appear to be false positives rather than duplicates.", + "summary": "The soft matches are mostly superficial title similarity across unrelated regressions (tokenizer, import, model-loading, quantization, and processor bugs). I would not merge any of these pairs as duplicates.", + "confidence": 0.81, + "canonical_issue_reason": "If a representative issue is needed, issue 43122 is the broadest tokenizer-regression report in the set (version-to-version behavior change), but it is only a loose thematic anchor, not a true duplicate umbrella.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue candidate: none of these issues is a clear representative of a duplicate group, since the cluster does not cohere around one bug.", + "best_issue_reason": "Issue 43122 is the least-specific representative among the tokenizer-related items, so it is the best single anchor here despite the cluster being too heterogeneous for a real canonical duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44263", - "right": "issue:45092", - "accept": false, - "reason": "Different bugs: a torch.split/indexer issue versus a remote-code/meta-initialization compatibility issue." - }, - { - "left": "issue:43232", - "right": "issue:46424", - "accept": false, - "reason": "Unrelated code paths: generation kwargs after sync_gpus versus cache max_cache_length handling." - }, - { - "left": "issue:43232", - "right": "issue:45834", - "accept": false, - "reason": "Generation internals and a Kosmos2.5 OCR index error are not the same bug." - }, - { - "left": "issue:43475", - "right": "issue:46305", - "accept": false, - "reason": "Different models and failures: missing SAM3 vision encoder attribute versus MoonshineStreaming mask off-by-one." - }, - { - "left": "issue:30990", - "right": "issue:38175", - "accept": false, - "reason": "Both are model-related, but one is loading hang and the other is zero probabilities in a different model." - }, { "left": "issue:46305", "right": "issue:46655", "accept": false, - "reason": "Both involve video processing, but the specific off-by-one mask bug and padding-calculation bug are different." + "reason": "Different bugs: sliding-window attention off-by-one vs video padding calculation." }, { "left": "issue:46209", "right": "issue:46431", "accept": false, - "reason": "FP8 inference on B200 is unrelated to missing chat template support in a processor." + "reason": "FP8 inference failure on B200 is unrelated to processor.save_pretrained missing files." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Different serialization problems: reading position_ids after buffer registration versus save_pretrained missing files." + "reason": "Reading a buffer after registration and missing save_pretrained files are different code paths." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Completely different logic errors: torch.split return values versus a missing ReLU in a DSA indexer." + "reason": "torch.split return handling is not the same as a missing ReLU in the DSA indexer." }, { - "left": "issue:45092", - "right": "issue:45357", + "left": "issue:44263", + "right": "issue:44373", "accept": false, - "reason": "Different regressions: meta-init incompatibility for InternVL2 versus wrong visual encoder keys on Qwen3.5 save_pretrained." + "reason": "A runtime indexing bug and a docstring typo are not duplicates." }, { - "left": "issue:44263", - "right": "issue:44373", + "left": "issue:45092", + "right": "issue:45357", "accept": false, - "reason": "Indexing bug versus a docstring typo; not the same underlying issue." + "reason": "Meta-init incompatibility for old InternVL2 and incorrect visual-encoder keys on save are different problems." }, { "left": "issue:42907", "right": "issue:46050", "accept": false, - "reason": "Both touch save/load of quantized models, but they affect different quantization paths and failure modes." + "reason": "Both touch quantized save/load, but the quantization stacks and failure modes differ." }, { "left": "issue:46369", "right": "issue:46909", "accept": false, - "reason": "A SAM3 LiteText memory overflow is unrelated to a broken documentation example for DIA." + "reason": "Memory overflow in sam3-litetext is unrelated to a broken documentation example." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "Different model families and symptoms: SigLIP2 zero probabilities versus Qwen3.5 MoE from_pretrained error." + "reason": "FP8 accuracy degradation and Qwen3.5 MoE from_pretrained failure are not the same bug." }, { "left": "issue:45381", "right": "issue:46305", "accept": false, - "reason": "Both are sequence-positioning bugs, but they occur in different models and mechanics; not the same code-path issue." - }, - { - "left": "issue:44483", - "right": "issue:45406", - "accept": false, - "reason": "An API request acceptance issue and a Gemma4Processor attribute error are unrelated." + "reason": "Both are indexing/positioning issues, but they affect different models and different mechanisms." }, { "left": "issue:44279", "right": "issue:46328", "accept": false, - "reason": "A dependency issue and a padding performance issue are not duplicates." + "reason": "A dependency problem and a SequenceFeatureExtractor.pad performance issue are unrelated." }, { - "left": "issue:22355", - "right": "issue:27957", + "left": "issue:44483", + "right": "issue:45406", "accept": false, - "reason": "Importing transformers.onnx and XLMRoberta Flash Attention 2 are unrelated problems." + "reason": "Chat/completions request rejection is unrelated to Gemma4Processor missing _tokenizer." }, { - "left": "issue:44829", - "right": "issue:46002", + "left": "issue:22355", + "right": "issue:27957", "accept": false, - "reason": "Both involve training quality, but one is Flash Attention 3 degeneracy and the other is NaNs after checkout/init; different causes." + "reason": "Missing transformers.onnx and XLMRoberta Flash Attention 2 are unrelated." }, { "left": "issue:43644", "right": "issue:46305", "accept": false, - "reason": "Non-persistent buffer junk in v5 is unrelated to a streaming mask off-by-one bug." + "reason": "Non-persistent buffer junk and sliding-window mask off-by-one are different failures." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "A missing rms_norm_eps in MLA layernorm is not the same as a remote-code/meta-init incompatibility." + "reason": "Missing rms_norm_eps precision issue is not the same as a meta-initialization incompatibility." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Both mention training instability, but the concrete causes and code paths differ." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "Different failures in different models: SigLIP2 zero probabilities versus Qwen3-VL-Moe bug." + "reason": "FP8 accuracy regression and a Qwen3-VL-Moe bug are not the same underlying issue." }, { "left": "issue:43408", "right": "issue:46402", "accept": false, - "reason": "A model-type/config mismatch warning and a missing text-encoder weight load are related only by SAM3 naming, not the same bug." + "reason": "A model-type warning and missing Sam3Model text encoder weights are different problems." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and likely different fixes." + "reason": "Missing AutoImageProcessor and missing Qwen2_5_VLForConditionalGeneration are separate import/export failures." }, { "left": "issue:42831", "right": "issue:45753", "accept": false, - "reason": "Zero probabilities in SigLIP2 and an infinite loop on a specific Qwen3.5 image are unrelated." + "reason": "FP8 accuracy issues and an infinite loop on a specific image are unrelated." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "A missing module import and tokenizer spacing behavior are different classes of issues." + "reason": "Missing onnx module and extra tokenizer spacing after added tokens are different bugs." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer bugs, but offset_mapping and add_special_tokens/BOS-EOS handling are distinct." + "reason": "Wrong offset_mapping and missing BOS/EOS insertion are distinct tokenizer behaviors." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline cache loading failure and wrong tokenizer decoder type are separate tokenizer issues." + "reason": "Offline cache loading failure and wrong tokenizer decoder type are related only at a broad tokenizer level, not the same defect." }, { "left": "issue:27957", "right": "issue:30990", "accept": false, - "reason": "Flash Attention 2 on XLMRoberta and a Sentence Transformers load hang are unrelated." + "reason": "Flash Attention 2 on XLMRoberta and Sentence Transformers loading stalls are unrelated." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Tokenizer offline cache loading and quantized model saving are different subsystems and bugs." + "reason": "Tokenizer offline-cache loading and quantized model saving are different code paths." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "A torch.compile attention-kernel divergence bug is unrelated to a docstring issue for position_ids." + "reason": "torch.compile attention divergence and a position_ids docstring typo are not duplicates." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related regressions, but one is a specific huggingface_hub symbol import and the other is a broader upgrade-time import breakage." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "A type-mismatch warning is unrelated to a memory overflow in model execution." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs after sync_gpus and missing auxiliary losses are different features." + }, + { + "left": "issue:42796", + "right": "issue:43122", + "accept": false, + "reason": "Tokenizer BPE-marker handling for one model and general version-to-version tokenization drift are not the same concrete bug." } ] }, @@ -15314,11 +15342,8 @@ "issue:39401", "issue:41628", "issue:42371", - "issue:42757", - "issue:42796", "issue:42907", "issue:43122", - "issue:43232", "issue:43408", "issue:43452", "issue:43475", @@ -15332,6 +15357,7 @@ "issue:44038", "issue:44060", "issue:44117", + "issue:44222", "issue:44263", "issue:44295", "issue:44297", @@ -15344,23 +15370,20 @@ "issue:44829", "issue:44945", "issue:44964", + "issue:45216", "issue:45278", "issue:45290", "issue:45381", - "issue:45593", + "issue:45478", "issue:45600", "issue:46002", "issue:46172", "issue:46326", - "issue:46369", + "issue:46614", "issue:46762", "issue:46909" ], "soft_pairs": [ - "issue:42757|issue:45278", - "issue:43408|issue:46369", - "issue:43232|issue:45593", - "issue:42796|issue:43122", "issue:43646|issue:46002", "issue:45381|issue:46909", "issue:43540|issue:46909", @@ -15373,182 +15396,162 @@ "issue:43576|issue:43784", "issue:42371|issue:44263", "issue:43756|issue:44060", - "issue:42907|issue:43937", "issue:46172|issue:46762", - "issue:43784|issue:45600", + "issue:42907|issue:43937", "issue:41628|issue:45278", + "issue:43784|issue:45600", "issue:44117|issue:44704", "issue:39401|issue:44521", "issue:44625|issue:44829", - "issue:44704|issue:45290", "issue:43408|issue:43475", + "issue:44704|issue:45290", "issue:43122|issue:44521", "issue:44360|issue:44485", "issue:43122|issue:46326", "issue:42907|issue:43452", "issue:44038|issue:44297", - "issue:22355|issue:29159" + "issue:22355|issue:29159", + "issue:45216|issue:45478", + "issue:44038|issue:44295", + "issue:39401|issue:46614", + "issue:44038|issue:44222" ], - "prepared_review_unit_hash": "56c0662aa75d428e617159bf550e359392fa0a9efef0e79dbb663c839f1b2ca0", + "prepared_review_unit_hash": "0124654e7956b7186de7f5b75008a097096d4b4106e0ad57be57e7d93f4cdf20", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36467, - "estimated_input_tokens": 9117, - "estimated_eval_tokens": 18490 + "serialized_chars": 35975, + "estimated_input_tokens": 8994, + "estimated_eval_tokens": 18244 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated Transformers regressions across imports, tokenizers, generation, multimodal, and model-specific paths. None of the soft pairs look like true duplicates of the same underlying bug/change, so I would keep them separate.", - "confidence": 0.95, - "canonical_issue_reason": "issue:45278 is the broadest and most representative triage anchor here: it captures a widespread import-regression theme, while the other issues are narrower model- or API-specific failures.", + "summary": "These items are a set of mostly unrelated issues across different models, tokenizers, and release regressions. I found no soft-edge pair that clearly represents the same underlying bug/change, so all candidate links are rejected.", + "confidence": 0.94, + "canonical_issue_reason": "No strong duplicate cluster emerges; the issues span distinct code paths and model families, with only loose thematic overlap in a few cases.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best global issue representative because it is the most umbrella-like user-facing regression and can serve as the cluster's main reference point.", + "best_issue_reason": "issue:44521 is the clearest standalone bug report: it has a specific, reproducible failure mode, broad impact for multimodal chat templating, and a focused fix target.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42757", - "right": "issue:45278", - "accept": false, - "reason": "Both are import failures, but one is a specific missing symbol from huggingface_hub and the other is a broad v5 import-regression report; not the same bug." - }, - { - "left": "issue:43408", - "right": "issue:46369", - "accept": false, - "reason": "Different SAM3 problems: a model-type/config warning versus a memory overflow when loading sam3-litetext." - }, - { - "left": "issue:43232", - "right": "issue:45593", - "accept": false, - "reason": "Generation kwarg handling and D-FINE auxiliary-loss behavior are unrelated code paths." - }, - { - "left": "issue:42796", - "right": "issue:43122", - "accept": false, - "reason": "Both are tokenizer-related, but one is a Ministral BPE-marker bug and the other is a version-to-version tokenization change; different root causes." - }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model initialization breakage is not the same as NaNs in classification heads after checkout/init." + "reason": "Different problems: custom model init breakage vs NaNs during classification head checkout/init." }, { "left": "issue:45381", "right": "issue:46909", "accept": false, - "reason": "Qwen2.5-VL video position IDs and a DIA documentation example are unrelated." + "reason": "Different subsystems: Qwen2.5-VL video position IDs vs a broken DIA documentation example." }, { "left": "issue:43540", "right": "issue:46909", "accept": false, - "reason": "Video-input processing in Qwen3OmniMoe is unrelated to a broken documentation example." + "reason": "Qwen3OmniMoe video-input processing failure is unrelated to a documentation example bug." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "Sentence Transformers loading hang and a MacOS bus error in CLIP are different failure modes and likely different fixes." + "reason": "Both are loading issues, but they affect different models and failure modes (stuck load vs bus error)." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Dequantized model saving and pipeline-parallel LLM output corruption are distinct runtime paths." + "reason": "Saving dequantized Ministrals/Devstrals is unrelated to pipeline-parallelism output corruption." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "Saving dequantized Ministral/Devstral models is unrelated to loading Phi-4 multimodal instruct." + "reason": "Different model families and different workflows: dequantized save vs model load failure." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "A tokenizer offset-mapping bug is not the same as an inability to load a specific model." + "reason": "Wrong offset mappings in Qwen3 tokenizer is unrelated to loading cjvt/sleng-bert." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python-version compatibility and pipeline-parallel output correctness are different issues." + "reason": "Python version compatibility regression is not the same bug as incorrect outputs under pipeline parallelism." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "Reading a registered position_ids buffer and tokenizer_config class mismatches are separate save/load bugs." + "reason": "One is a position_ids buffer access error; the other is a tokenizer_class mismatch on save." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "A broken CLI env command and an import-time NameError in sentence-transformers are unrelated." + "reason": "Broken env command and a sentence-transformers import NameError are different failure paths." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 settings API guidance and a torch.split return-value issue in GlmMoeDsaIndexer do not share a code path." + "reason": "TF32 settings warning and a torch.split return-value issue are unrelated." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Smollm3 RoPE-layer dropping and Qwen3-Next tied-weights warnings are different model implementation bugs." - }, - { - "left": "issue:42907", - "right": "issue:43937", - "accept": false, - "reason": "Model-saving failure and GLM-5 GenerationConfig validation are unrelated." + "reason": "Both involve model internals, but Smollm3 RoPE-layer mismatch and Qwen3-Next tied-weights warning are distinct bugs." }, { "left": "issue:46172", "right": "issue:46762", "accept": false, - "reason": "DeepSeek-V4 dtype/transpose behavior and Minimax MSA numerical differences are different attention/math issues." + "reason": "Different numerical issues in different models/attention implementations; not the same code-path problem." }, { - "left": "issue:43784", - "right": "issue:45600", + "left": "issue:42907", + "right": "issue:43937", "accept": false, - "reason": "Importing sentence-transformers with latest transformers and stale auto_mappings entries for Sam3LiteText configs are separate breakages." + "reason": "Dequantized save failure is unrelated to invalid GenerationConfig handling." }, { "left": "issue:41628", "right": "issue:45278", "accept": false, - "reason": "A specific AutoImageProcessor import error is not the same underlying bug as a broad set of import errors after upgrade." + "reason": "A single missing import is not the same as a broad post-upgrade import-regression report." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Sentence-transformers import failure and stale auto_mappings references are separate import-time issues." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "Tokenizer mapping fallback behavior and AutoProcessor cached_file kwarg propagation are different failure points." + "reason": "Tokenizer mapping None-handling and AutoProcessor kwarg forwarding are different bugs." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Wrong tokenizer offset mappings and zero assistant masks in multimodal templates are distinct tokenizer/template bugs." + "reason": "Tokenizer offset mapping bug is unrelated to multimodal assistant mask generation." }, { "left": "issue:44625", @@ -15557,40 +15560,76 @@ "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training degeneration are unrelated." }, { - "left": "issue:44704", - "right": "issue:45290", + "left": "issue:43408", + "right": "issue:43475", "accept": false, - "reason": "AutoProcessor cached_file kwargs and apply_chat_template tool-call crashes are different API paths." + "reason": "Both are SAM3-related, but model-type warning and missing attribute error are distinct issues." }, { - "left": "issue:43408", - "right": "issue:43475", + "left": "issue:44704", + "right": "issue:45290", "accept": false, - "reason": "Both touch SAM3, but one is a config/model-type mismatch warning and the other is a missing attribute in vision encoder output." + "reason": "cached_file kwargs forwarding and chat-template tool-call crashes are different code paths." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Version-related tokenization differences and multimodal assistant-mask generation are not the same bug." + "reason": "Tokenization differences across versions are not the same as multimodal assistant_mask failures." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "DSA indexer activation choice and GLM-5 RoPE implementation are unrelated model internals." + "reason": "DSA indexer ReLU discussion and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Both concern tokenizer behavior, but they are separate bugs in different tokenizer stacks." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "GGUF loading breakage is unrelated to dequantized save failures." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe loading and tokenizer_class metadata mismatches are separate problems." + "reason": "Qwen3-VL-Moe version bug and tokenizer_class save mismatch are unrelated." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "Missing transformers.onnx module and slow/fast tokenizer inconsistency are unrelated issues." + "reason": "No module import failure and slow/fast tokenizer inconsistency are different problems." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 and version regressions, but one is save_pretrained checkpoint correctness and the other is from_pretrained loading failure; not clearly the same concrete bug." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe bug and position_ids buffer access error do not share the same underlying failure." + }, + { + "left": "issue:39401", + "right": "issue:46614", + "accept": false, + "reason": "Qwen3 tokenizer offset mapping and stale phonemizer backend state are unrelated tokenizer issues." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Different Qwen3-VL-Moe issue vs FP8 save_pretrained moe bug; same broad area, different concrete failures." } ] }, @@ -15618,7 +15657,6 @@ "issue:29127", "issue:36296", "issue:38617", - "issue:39401", "issue:42175", "issue:42371", "issue:42673", @@ -15630,26 +15668,27 @@ "issue:43278", "issue:43493", "issue:43611", + "issue:43618", "issue:43824", + "issue:43825", "issue:43874", "issue:43937", "issue:43976", "issue:44038", + "issue:44075", "issue:44188", - "issue:44222", - "issue:44295", "issue:44351", "issue:44355", "issue:44485", "issue:44610", "issue:44625", "issue:44779", + "issue:44936", "issue:44938", "issue:44945", "issue:45072", "issue:45083", "issue:45200", - "issue:45216", "issue:45278", "issue:45357", "issue:45446", @@ -15659,16 +15698,13 @@ "issue:46231", "issue:46328", "issue:46614", + "issue:46829", "issue:46877", "issue:46909" ], "soft_pairs": [ - "issue:45216|issue:45478", - "issue:44038|issue:44295", - "issue:39401|issue:46614", - "issue:44038|issue:44222", - "issue:43116|issue:45200", "issue:43824|issue:44351", + "issue:43116|issue:45200", "issue:43122|issue:44779", "issue:43278|issue:44945", "issue:36296|issue:43116", @@ -15680,228 +15716,225 @@ "issue:46231|issue:46877", "issue:43937|issue:44485", "issue:42898|issue:46614", - "issue:43066|issue:44779", "issue:38617|issue:45278", + "issue:43066|issue:44779", "issue:43611|issue:46002", "issue:43976|issue:45446", "issue:44038|issue:45478", - "issue:43116|issue:43493", "issue:42673|issue:45072", + "issue:43116|issue:43493", "issue:45357|issue:45478", "issue:44355|issue:46231", "issue:44625|issue:46002", "issue:42371|issue:43874", + "issue:45083|issue:45446", "issue:44610|issue:46909", - "issue:45083|issue:45446" + "issue:44075|issue:44936", + "issue:43618|issue:46829", + "issue:43116|issue:43825" ], - "prepared_review_unit_hash": "338db466e34638615685215b17b2d4de5638929da9506f33cf99ecb37434693c", + "prepared_review_unit_hash": "33dc85a766262b8c5bb912822567482f77ee144f650f3d9823327f74eef09c73", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 35707, - "estimated_input_tokens": 8927, - "estimated_eval_tokens": 18110 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35984, + "estimated_input_tokens": 8996, + "estimated_eval_tokens": 18248 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: the soft pairs mostly share broad themes like tokenizers, imports, or Qwen3.5, but they do not look like the same concrete bug/change. I would not collapse any of the provided soft edges.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "This cluster is heterogeneous: most pairs share only broad vocabulary (v5 regressions, imports, tokenizers, multimodal models) but do not describe the same concrete bug or change. No soft edge looks safe to merge as a duplicate.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45278 is the broadest high-signal report in the set: a v5 upgrade causing multiple import failures. It is the closest thing to a hub for the cluster, though the cluster as a whole is not a true duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45278 is the best representative of the overall theme of upgrade/regression breakage. It is still too broad to unify the other items as duplicates, but it is the most central issue here.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45216", - "right": "issue:45478", - "accept": false, - "reason": "Both involve Qwen3.5 save/load behavior, but one is a bad save_pretrained checkpoint and the other is a from_pretrained load error for Qwen3.5-MoE; different failing paths." - }, - { - "left": "issue:44038", - "right": "issue:44295", - "accept": false, - "reason": "Different subsystems and symptoms: Qwen3-VL-Moe bug vs a position_ids buffer access error." - }, - { - "left": "issue:39401", - "right": "issue:46614", - "accept": false, - "reason": "Both are tokenizer-related, but they affect different models and different code paths (offset mapping vs phonemizer backend state)." - }, - { - "left": "issue:44038", - "right": "issue:44222", + "left": "issue:43824", + "right": "issue:44351", "accept": false, - "reason": "Same broad MoE area, but one is a Qwen3-VL-Moe issue and the other is FP8 save_pretrained behavior; not the same bug." + "reason": "Both are import errors, but for different missing symbols and likely different breakpoints; not the same bug." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Different features: example script multi-label outputs vs Gemma 4 text-only fine-tuning defaults." - }, - { - "left": "issue:43824", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but for different missing symbols from different modules." + "reason": "Classification example output vs Gemma multimodal token ID handling are unrelated code paths." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they concern different models and different symptoms." + "reason": "Both are tokenizer regressions in v5, but they affect different models and symptoms; too broad to merge." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "One is an eval-time dtype change, the other is incorrect output under pipeline parallelism; no shared concrete failure." + "reason": "Embedding dtype drift during eval and pipeline-parallel incorrect outputs are different failure modes." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor parallel training bug and classification example output failure are unrelated." + "reason": "Tensor parallel training bug and example script multi-label empties are unrelated." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they involve different removed/renamed symbols and unrelated call sites." + "reason": "Different import sources and missing names; same symptom class, not same underlying issue." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Missing transformers.onnx and a LayoutLMv3 error-message clarity issue are unrelated." + "reason": "One is a missing module import, the other is an error-message clarity complaint in LayoutLMv3." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Package/backend install coverage and Python 3.14 load failure are different problems." + "reason": "Dependency/backend packaging issue vs Python 3.14 load failure; no shared concrete code path." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "SigLIP2 implementation discrepancy and a Qwen3.5 image infinite-loop bug are unrelated." + "reason": "Different multimodal model regressions with different behavior and triggers." }, { "left": "issue:44188", "right": "issue:46328", "accept": false, - "reason": "torch.compile attention kernel divergence is unrelated to a padding performance issue." + "reason": "Attention-kernel divergence and feature-extractor padding performance are unrelated." }, { "left": "issue:46231", "right": "issue:46877", "accept": false, - "reason": "One is a docs parameter-name mismatch; the other is import-time lint output pollution." + "reason": "Both touch docs, but one is a parameter-name doc bug and the other is import-time lint output." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "A bad GenerationConfig validation report is not the same as a RoPE implementation discussion." + "reason": "Both mention GLM-5, but one is GenerationConfig validation and the other is RoPE implementation; different bugs." }, { "left": "issue:42898", "right": "issue:46614", "accept": false, - "reason": "v5 tokenization cleanup changes and stale phonemizer language state are different bugs." + "reason": "Tokenizer cleanup behavior and phonemizer backend caching are different subsystems." }, { - "left": "issue:43066", - "right": "issue:44779", + "left": "issue:38617", + "right": "issue:45278", "accept": false, - "reason": "Different tokenizer regressions affecting different models and behaviors." + "reason": "Specific missing import vs broad post-upgrade import failures; 45278 is broader but not the same defect." }, { - "left": "issue:38617", - "right": "issue:45278", + "left": "issue:43066", + "right": "issue:44779", "accept": false, - "reason": "Specific layer_type_validation import failure is not the same as a broad report of many import errors after upgrade." + "reason": "Both are tokenizer-related regressions, but the concrete symptoms and affected paths differ." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Base-model-prefix loading regression and NaNs in classification heads are unrelated." + "reason": "Model loading with base_model_prefix and NaNs in classification heads are different problems." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version support problem and a PyTorch version check bug are different issues." + "reason": "Python version compatibility failure and an AuxRequest version-check bug are unrelated." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Both mention Qwen3.5/MoE, but they describe different models and different load-time failures." + "reason": "Same model family, but different versions and different loading failures; not one mergeable fix." }, { - "left": "issue:43116", - "right": "issue:43493", + "left": "issue:42673", + "right": "issue:45072", "accept": false, - "reason": "Classification script output bug and SigLIP2 discrepancy are unrelated." + "reason": "VRAM leak in multi-threaded dataloading is unrelated to bfloat16 dtype mismatches in CI." }, { - "left": "issue:42673", - "right": "issue:45072", + "left": "issue:43116", + "right": "issue:43493", "accept": false, - "reason": "VRAM leak with multiple dataloader threads is unrelated to dtype mismatch failures in CI." + "reason": "Example script classification behavior and SigLIP2 implementation discrepancy do not share a bug." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen3.5 save/load regressions, but one is about wrong visual encoder keys on save_pretrained and the other is a from_pretrained error for MoE." + "reason": "Both involve Qwen3.5/MoE regressions, but one is save_pretrained key handling and the other is from_pretrained loading." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "Compiled Python file errors and a documentation parameter mismatch are unrelated." + "reason": "Compiled-file import errors and a documentation parameter mismatch are unrelated." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "num_labels propagation and NaNs in classification heads are different bugs." - }, - { - "left": "issue:45083", - "right": "issue:45446", - "accept": false, - "reason": "Feature-length helper behavior and an AuxRequest version-check bug do not share the same code path." + "reason": "Config propagation for num_labels and NaNs on init/checkpoint are different failure modes." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 backend settings and a missing image-processor method are unrelated." + "reason": "TF32 policy messaging and a missing image-patch method are unrelated." }, { - "left": "issue:44610", + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Different helper/version-check bugs in different modules; no shared code path." + }, + { + "left": "issue:44610", "right": "issue:46909", "accept": false, - "reason": "Processor/model input-size mismatch and a broken documentation example are not the same underlying issue." + "reason": "Processor input-size mismatch and a broken documentation example for DIA are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer args not applied and trainer.evaluate() failing after train are separate training issues." + }, + { + "left": "issue:43618", + "right": "issue:46829", + "accept": false, + "reason": "Both involve output fields, but one is CLIP attentions missing and the other is VideoPrism hidden_states misassigned." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example script multi-label output bug and pipeline translation-task messaging are unrelated." } ] }, @@ -15933,15 +15966,13 @@ "issue:42796", "issue:43064", "issue:43066", - "issue:43116", "issue:43441", - "issue:43618", + "issue:43450", "issue:43673", - "issue:43825", "issue:43827", "issue:43937", + "issue:43986", "issue:43992", - "issue:44075", "issue:44117", "issue:44263", "issue:44291", @@ -15950,11 +15981,13 @@ "issue:44466", "issue:44704", "issue:44779", - "issue:44936", + "issue:44829", "issue:44977", "issue:44995", "issue:45071", + "issue:45081", "issue:45092", + "issue:45216", "issue:45310", "issue:45341", "issue:45405", @@ -15965,26 +15998,23 @@ "issue:46002", "issue:46077", "issue:46328", + "issue:46512", "issue:46614", "issue:46640", "issue:46655", "issue:46697", "issue:46710", "issue:46752", - "issue:46762", - "issue:46829" + "issue:46762" ], "soft_pairs": [ - "issue:44075|issue:44936", - "issue:43618|issue:46829", - "issue:43116|issue:43825", "issue:43066|issue:46710", "issue:43992|issue:44704", "issue:46614|issue:46752", "issue:44995|issue:46697", - "issue:28348|issue:46640", "issue:44263|issue:46762", "issue:45092|issue:46762", + "issue:28348|issue:46640", "issue:43827|issue:45446", "issue:33453|issue:39401", "issue:43064|issue:43673", @@ -16002,204 +16032,221 @@ "issue:39401|issue:44779", "issue:43937|issue:44977", "issue:44410|issue:45310", - "issue:45341|issue:45561" + "issue:45341|issue:45561", + "issue:45071|issue:46512", + "issue:45341|issue:46328", + "issue:43450|issue:43986", + "issue:44829|issue:45216", + "issue:44117|issue:45081" ], - "prepared_review_unit_hash": "006f4f682d2a7883f21740a0443b5fb3ee352aa2f4d7b71ebaddb1c1d5c5e224", + "prepared_review_unit_hash": "1130b9a068554166a69152e664ab1f9ecdb04df44b7c6a6dcc7a779a0aec9b29", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 35814, - "estimated_input_tokens": 8954, - "estimated_eval_tokens": 18164 + "soft_pair_count": 30, + "serialized_chars": 36667, + "estimated_input_tokens": 9167, + "estimated_eval_tokens": 18590 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose similarity group of mostly unrelated Transformers v5 issues (tokenizers, generation, attention backends, model loading, and testing utils). None of the soft pairs look like the same underlying bug or change, so no duplicate merges should be made.", - "confidence": 0.91, - "canonical_issue_reason": "No pair is close enough to be a true duplicate; if one issue must be kept as the anchor, issue 42673 is one of the clearer, well-scoped runtime bugs in the set, but it is not actually representative of the others.", + "summary": "This cluster is broad and mostly heterogeneous: tokenizer loading/format regressions, attention-backend issues, video/processor bugs, and a few training/generation failures. I don\u2019t see any soft pair that clearly describes the exact same underlying bug, so I would keep them separate. The most umbrella-like issue is the generic tokenizer-loading regression.", + "confidence": 0.57, + "canonical_issue_reason": "issue:33453 is the broadest, most umbrella-style report in the set (\u201cRegression in tokenizer loading\u201d) and best matches the cluster\u2019s repeated tokenizer-related regressions without overfitting to one model.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 42673 is the strongest standalone issue in the cluster: concrete repro, clear impact (VRAM leak), and good discussion activity. Still, it is not a duplicate target for the other items.", + "best_issue_reason": "issue:33453 is the best representative issue because it is generic enough to cover several tokenizer-loading variants in the cluster, while the other issues are too model- or symptom-specific.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44075", - "right": "issue:44936", - "accept": false, - "reason": "Optimizer-argument bug vs trainer post-train/evaluate failure; different code paths and symptoms." - }, - { - "left": "issue:43618", - "right": "issue:46829", - "accept": false, - "reason": "Both are output-field bugs, but for different models and different output objects; not the same concrete fix." - }, - { - "left": "issue:43116", - "right": "issue:43825", - "accept": false, - "reason": "Example-script multi-label result bug vs pipeline translation error-message issue; unrelated." - }, { "left": "issue:43066", "right": "issue:46710", "accept": false, - "reason": "Tokenizer decoder-type regression vs accuracy regression on a specific model; too broad and not the same bug." + "reason": "Both are v5 regressions, but one is tokenizer decoder typing and the other is a model accuracy regression; different failures and code paths." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Missing weight in UMT5 encoder loading vs AutoProcessor kwargs forwarding; different loading layers." + "reason": "UMT5 weight loading and AutoProcessor kwargs forwarding are unrelated loader issues." }, { "left": "issue:46614", "right": "issue:46752", "accept": false, - "reason": "Phonemizer backend cache bug vs empty-list chat template IndexError; unrelated features." + "reason": "Stale phonemizer backend vs empty-list chat templating crash are different subsystems." }, { "left": "issue:44995", "right": "issue:46697", "accept": false, - "reason": "GLM-MoE stale indexer cache crash vs Swin2SR padding overrun; different subsystems." - }, - { - "left": "issue:28348", - "right": "issue:46640", - "accept": false, - "reason": "Feature request for GPT-2 flash-attn support vs a generic flash-attention issue; not the same change." + "reason": "GlmMoeDsa cache corruption and image over-padding are unrelated bugs." }, { "left": "issue:44263", "right": "issue:46762", "accept": false, - "reason": "torch.split return handling in one model vs numerical-difference attention discrepancy in another; different bugs." + "reason": "Different attention/model numerical issues; not the same concrete bug." }, { "left": "issue:45092", "right": "issue:46762", "accept": false, - "reason": "Remote-code/meta-init incompatibility vs attention numerical mismatch; unrelated." + "reason": "Meta-init incompatibility vs attention numerical drift are distinct problems." + }, + { + "left": "issue:28348", + "right": "issue:46640", + "accept": false, + "reason": "A flash-attention support request and a generic flash-attention issue are not the same bug." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs referencing removed pipeline API vs PyTorch version check in flex_attention; unrelated." + "reason": "Docs cleanup and a version-check bug in flex_attention are unrelated." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Both are tokenizer regressions, but one is loading while the other is offset_mapping for Qwen3; not the same concrete issue." + "reason": "Generic tokenizer-loading regression vs Qwen3 offset_mapping correctness; related area but different bug." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Distributed optimizer-state corruption vs generation cache missing during chunked prefill; different training/inference paths." + "reason": "Optimizer-state corruption under FSDP/PEFT is unrelated to missing generation cache in chunked prefill." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "Vitpose breakage vs Ministral flash-attention failure; unrelated models and failures." + "reason": "Broken vitpose models and Ministral flash-attention failures are different model-specific issues." }, { "left": "issue:42673", "right": "issue:42796", "accept": false, - "reason": "VRAM leak in multi-threaded Qwen3 loading vs BPE-marker tokenizer bug; unrelated." + "reason": "VRAM leak in threaded loading vs tokenizer BPE-marker handling are unrelated." }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "Invalid GenerationConfig vs stale num_return_sequences warning; different generation bugs." + "reason": "Invalid GenerationConfig and a stale warning about num_return_sequences are different generation bugs." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Chunked-prefill cache regression vs sequence feature extractor padding performance issue; unrelated." + "reason": "Generation cache missing and SequenceFeatureExtractor padding performance are unrelated." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "Unexpected init arg during meta-init vs PretrainedConfig type-checking breakage; both v5-related but distinct." + "reason": "Unexpected init_empty_weights argument vs PretrainedConfig type-checking breakage are different v5 regressions." }, { "left": "issue:42673", "right": "issue:46655", "accept": false, - "reason": "VRAM leak in dataloader-thread usage vs incorrect video padding; unrelated." + "reason": "VRAM leak and video padding are unrelated." }, { "left": "issue:42222", "right": "issue:46752", "accept": false, - "reason": "Vitpose model breakage vs empty-list chat template crash; different APIs." + "reason": "Vitpose breakage and empty-list chat template crash do not share the same code path." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "Dependency version bump problem vs PyTorch version gate bug; not duplicates." + "reason": "PEFT version bump and flex_attention import version check are unrelated." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "Tokenizer mapping None handling vs flash_attention sink-less AttributeError; different code paths." + "reason": "Tokenizer mapping None and flash_attention s_aux=None crash are different defects." }, { "left": "issue:44466", "right": "issue:46002", "accept": false, - "reason": "State serialization inconsistency vs NaNs after checkout/init; not the same underlying bug." + "reason": "lm_head serialization inconsistency and NaNs in classification heads are not the same issue." }, { "left": "issue:44380", "right": "issue:46077", "accept": false, - "reason": "GPT-2 attention scaling ignored vs Gemma4 bidirectional mask bug; same area, different fix." + "reason": "GPT2 attention scaling ignored and Gemma4 bidirectional masks are separate attention-backend bugs." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Qwen3 offset_mapping bug vs DeepSeek tokenizer regression; both tokenizer issues but different models and likely different root causes." + "reason": "Both are tokenizer regressions, but they affect different models and symptoms; not clearly the same bug." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "Invalid GenerationConfig vs flash-attention generation failure in Qwen3.5; too different to merge." + "reason": "GenerationConfig validation and Qwen3.5 flash-attention generation failure are different failures." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing Qwen3next projections vs Qwen3.5 MoE from_pretrained error; related family but not the same bug." + "reason": "Missing Qwen3Next projections vs Qwen3.5 MoE from_pretrained error are related only by model family, not the same concrete bug." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Generic testing_utils bug vs xdist race on captured_info.txt; distinct defects." + "reason": "A testing_utils bug and a pytest-xdist race are not the same issue." + }, + { + "left": "issue:45071", + "right": "issue:46512", + "accept": false, + "reason": "Type-checking breakage and stale config.dtype are distinct issues." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "Testing-utils bug and feature-extractor padding performance are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video shape bug vs missing-torchvision AutoProcessor crash are different video-processing failures." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "flash_attention_3 training degeneration and Qwen3.5 save_pretrained regression are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping fallback and Mistral regex patch crash are separate tokenizer-loading bugs." } ] }, @@ -16234,50 +16281,44 @@ "issue:43122", "issue:43208", "issue:43408", - "issue:43450", + "issue:43673", "issue:43742", + "issue:43874", + "issue:43931", "issue:43937", - "issue:43986", "issue:44075", - "issue:44117", "issue:44220", "issue:44261", "issue:44263", + "issue:44360", "issue:44410", "issue:44483", "issue:44484", "issue:44485", "issue:44493", "issue:44496", - "issue:44829", "issue:44877", "issue:44933", "issue:45071", - "issue:45081", - "issue:45216", "issue:45278", "issue:45310", "issue:45341", "issue:45362", "issue:45464", "issue:45478", + "issue:45542", "issue:45656", + "issue:46172", "issue:46178", "issue:46209", "issue:46231", "issue:46298", "issue:46328", - "issue:46512", "issue:46617", "issue:46762", "issue:46894" ], "soft_pairs": [ - "issue:45071|issue:46512", - "issue:45341|issue:46328", - "issue:43450|issue:43986", - "issue:44829|issue:45216", - "issue:44117|issue:45081", "issue:43408|issue:43742", "issue:42617|issue:45341", "issue:43122|issue:46298", @@ -16299,198 +16340,231 @@ "issue:43066|issue:43122", "issue:44263|issue:44485", "issue:44483|issue:45464", - "issue:43097|issue:44493" + "issue:43097|issue:44493", + "issue:43673|issue:45341", + "issue:43931|issue:44410", + "issue:42222|issue:45542", + "issue:41628|issue:45071", + "issue:45071|issue:45542", + "issue:44484|issue:45478", + "issue:43874|issue:44261", + "issue:44263|issue:46172", + "issue:44360|issue:46172" ], - "prepared_review_unit_hash": "4997c7059ed1f3d85210be100889bc1d0af1dba4eaef60f4f0ec4c098ae84087", + "prepared_review_unit_hash": "f26fff9d5081876e11ef5b25d881d5d0b97b6ad12eb48b684f82e7be0eeb7bb8", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35357, - "estimated_input_tokens": 8840, - "estimated_eval_tokens": 17936 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 35722, + "estimated_input_tokens": 8931, + "estimated_eval_tokens": 18118 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is dominated by superficially similar Transformers bug reports, but the soft pairs mostly describe different failure modes, models, or code paths. I found no convincing duplicates to merge.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "The cluster is highly heterogeneous: mostly unrelated import/model/config/tokenization regressions. None of the soft pairs looks like the same concrete bug/change, so I would not merge any of them.", + "confidence": 0.9, + "canonical_issue_reason": "If one issue must stand in for the cluster, 45278 is the broadest regression report ('many import errors after update') and best reflects the most general user-facing problem among these otherwise unrelated issues.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "45278 is the most representative umbrella issue because it captures a broad cross-version regression with multiple import failures, whereas the others are narrower model- or subsystem-specific bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:45071", - "right": "issue:46512", - "accept": false, - "reason": "Both involve config/type handling, but one is about PreTrainedConfig type checking in v5.4 while the other is about dtype state after from_config; different bugs and fixes." - }, - { - "left": "issue:45341", - "right": "issue:46328", - "accept": false, - "reason": "Testing utility doc/behavior bug versus a SequenceFeatureExtractor.pad performance issue; no shared code-path problem." - }, - { - "left": "issue:43450", - "right": "issue:43986", - "accept": false, - "reason": "One reports wrong batched video processor output shape, the other a missing torchvision dependency crash in AutoProcessor loading; different failure modes." - }, - { - "left": "issue:44829", - "right": "issue:45216", - "accept": false, - "reason": "Flash Attention 3 training degeneration and Qwen3.5 save_pretrained checkpoint corruption are unrelated defects." - }, - { - "left": "issue:44117", - "right": "issue:45081", - "accept": false, - "reason": "Tokenizer mapping None handling versus a Mistral regex patch crash; both touch loading, but not the same underlying bug." - }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "A model-class warning for SAM3 tracker/video mismatch is unrelated to a MobileLLM key error on load." + "reason": "Different models and failure modes: sam3 tracker/type warning vs MobileLLM key error." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "3d_parallel.py execution failure and a testing_utils bug do not describe the same change or code path." + "reason": "Unrelated topics: running 3d_parallel.py vs a small testing_utils bug." }, { "left": "issue:43122", "right": "issue:46298", "accept": false, - "reason": "Tokenizer behavior regression versus a noisy logging message; no shared defect." + "reason": "Tokenizer output regression vs noisy logging; no shared code-path bug." }, { "left": "issue:45362", "right": "issue:46328", "accept": false, - "reason": "Qwen3.5 chat crash and a padding performance issue are unrelated." + "reason": "Chat crash on Qwen3.5 vs feature-extractor padding performance; unrelated." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and different modules." + "reason": "Both are import errors, but for different missing names/modules and likely different fixes." }, { "left": "issue:42222", "right": "issue:46178", "accept": false, - "reason": "VitPose model breakage and an unsupported backend mapping error are not the same bug." + "reason": "VitPose model loading breakage vs backend mapping error for tensorflow_text; different defects." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Strict config loading and a documentation parameter-name mismatch are unrelated." + "reason": "Strict config loading vs documentation parameter mismatch; not the same issue." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "max_shard_size default behavior and a Qwen3.5 MoE from_pretrained failure are different issues." + "reason": "Different problems in different code paths: shard size behavior vs Qwen3.5 MoE from_pretrained error." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "Invalid GenerationConfig handling versus missing model_type config.json loading are distinct code paths." + "reason": "GenerationConfig validation for GLM-5 vs unrecognized model config; unrelated failures." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 API guidance and MLA rms_norm_eps precision behavior are unrelated." + "reason": "TF32 settings API guidance vs MLA precision/config bug; not the same underlying defect." }, { "left": "issue:39401", "right": "issue:46209", "accept": false, - "reason": "Tokenizer offset_mapping bug and FP8 inference failure on B200 are different subsystems." + "reason": "Tokenizer offset mapping bug vs FP8 inference on B200; different subsystems entirely." }, { "left": "issue:44261", "right": "issue:46762", "accept": false, - "reason": "One is a config epsilon mismatch causing precision drift; the other is a numerical difference in Minimax attention implementation." + "reason": "Both mention numerical differences, but one is a config epsilon issue and the other is a Minimax attention reference mismatch." }, { "left": "issue:42371", "right": "issue:46617", "accept": false, - "reason": "API settings guidance versus trust_remote_code cache FileNotFoundError are unrelated." + "reason": "TF32 backend guidance vs cache FileNotFoundError for trust_remote_code; unrelated." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Missing qwen3next projections and Qwen3.5 MoE from_pretrained error are different model-loading/model-definition problems." + "reason": "Different Qwen3-related loading errors: missing projections vs from_pretrained failure." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "A missing image_utils import and broad import errors after upgrade are related only at a very high level, not the same concrete bug." + "reason": "Shared import-error theme is too broad; one is a specific nonexistent import, the other a wide upgrade regression." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "An import-name regression and a fbank feature extraction bug are unrelated." + "reason": "ImportError for configuration_utils vs audio feature extraction issue; no same bug." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "SGD argument handling and duplicated optimizer.step under deepspeed are different optimizer issues." + "reason": "SGD args ignored and optimizer step called twice are different optimizer bugs." }, { "left": "issue:43208", "right": "issue:46894", "accept": false, - "reason": "xLSTM training blockers and a CohereASR double-shift loss bug do not share the same code path." + "reason": "xLSTM training blockers vs CohereASR double-shift loss bug; not the same concrete change." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both involve tokenization changes in v5, but one is wrong decoder type and the other is differing tokenization output; too broad to be the same bug." + "reason": "Both tokenizer-related, but one is decoder type mismatch and the other is tokenization behavior change; not one bug." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "A torch.split return-value bug in an indexer is not the same as a GLM-5 RoPE implementation discussion." + "reason": "torch.split return values vs GLM-5 RoPE implementation; unrelated code paths." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Generic chat/completions request rejection and streaming inference failure on a specific Qwen model are different endpoint/model issues." + "reason": "General chat/completions rejection in v5.3 vs streaming inference failure on Qwen3.5; not the same defect." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Removal of tie_embeddings_and_encoder_decoder and unexpected position-id keys are unrelated regression symptoms." + "reason": "Removed tie_embeddings flag vs unexpected position-id keys; different regressions." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing in chunked_prefill vs testing_utils bug; unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Weight-shape mismatch when loading Qwen3-VL vs missing projections in qwen3next; different model-loading problems." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Model breakage in vitpose vs TensorFlow backend initialization error; not the same underlying bug." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Different import/type issues: missing AutoImageProcessor vs PreTrainedConfig type checking regression." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Type checking regression vs undefined tf backend; unrelated." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Same broad model family is not enough; max_shard_size behavior and MoE from_pretrained failure are different bugs." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different multimodal/image-processor attribute bug vs MLA epsilon precision issue." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "Different GLM/DeepSeek implementation bugs with no shared failing code-path." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "Missing ReLU in DSA indexer vs tensor transpose bug in deepseek-v4; unrelated." } ] }, @@ -16516,9 +16590,8 @@ "nodes": [ "issue:27957", "issue:38617", - "issue:41628", - "issue:42222", "issue:42757", + "issue:42890", "issue:43012", "issue:43064", "issue:43097", @@ -16529,28 +16602,27 @@ "issue:43531", "issue:43577", "issue:43673", + "issue:43873", "issue:43874", "issue:43931", "issue:43986", "issue:44038", + "issue:44062", "issue:44117", "issue:44261", "issue:44263", "issue:44315", - "issue:44360", + "issue:44351", "issue:44410", "issue:44484", "issue:44485", "issue:44829", "issue:44987", "issue:45003", - "issue:45071", "issue:45341", "issue:45372", "issue:45375", "issue:45440", - "issue:45478", - "issue:45542", "issue:45588", "issue:45753", "issue:46050", @@ -16560,20 +16632,13 @@ "issue:46369", "issue:46655", "issue:46664", + "issue:46710", "issue:46752", + "issue:46821", "issue:46897" ], "soft_pairs": [ - "issue:43673|issue:45341", - "issue:43931|issue:44410", - "issue:42222|issue:45542", - "issue:41628|issue:45071", - "issue:45071|issue:45542", - "issue:44484|issue:45478", "issue:27957|issue:43352", - "issue:43874|issue:44261", - "issue:44263|issue:46172", - "issue:44360|issue:46172", "issue:43931|issue:44038", "issue:43441|issue:44829", "issue:43012|issue:43577", @@ -16586,8 +16651,8 @@ "issue:43531|issue:45341", "issue:43874|issue:44263", "issue:45440|issue:46172", - "issue:43097|issue:45588", "issue:43673|issue:46123", + "issue:43097|issue:45588", "issue:44410|issue:45375", "issue:43986|issue:45372", "issue:44038|issue:44410", @@ -16597,234 +16662,222 @@ "issue:43064|issue:44315", "issue:43986|issue:44484", "issue:46655|issue:46664", - "issue:46298|issue:46752" + "issue:46298|issue:46752", + "issue:46172|issue:46710", + "issue:42890|issue:46298", + "issue:44062|issue:44117", + "issue:43097|issue:43873", + "issue:38617|issue:44351", + "issue:43012|issue:46821" ], - "prepared_review_unit_hash": "4a037742314b2ec4c5d3b65ff19ca9ffa70189e2b9fb2410235d7bf1503aa61a", + "prepared_review_unit_hash": "ac7cfc352ec43947854af3ca5c1fd1c6d109daf2797ad95f9e2be6e1a8672889", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37634, - "estimated_input_tokens": 9409, - "estimated_eval_tokens": 19074 + "soft_pair_count": 31, + "serialized_chars": 36860, + "estimated_input_tokens": 9215, + "estimated_eval_tokens": 18686 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of superficially similar Transformers regressions (imports, model loading, multimodal processors, training bugs) rather than one duplicate bug. I would not merge any of the soft-edge pairs.", - "confidence": 0.2, - "canonical_issue_reason": "No single canonical issue stands out: the items span different code paths and bug classes, so they do not collapse into one underlying defect.", - "canonical_pr_reason": "No pull requests are present in this cluster.", - "best_issue_reason": "issue:43986 is the clearest representative user-facing regression here: a specific crash in AutoProcessor video-model loading with a clear dependency-triggered code path.", - "best_pr_reason": "No pull requests are present in this cluster.", + "summary": "The cluster is highly heterogeneous; the soft edges are mostly false positives across unrelated model families and bug classes. No PRs are present.", + "confidence": 0.18, + "canonical_issue_reason": "issue:43097 is the broadest, most representative release-regression report in the set and has inbound references; the rest are mostly narrow model-specific failures.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the clearest general user-facing regression and the least specialized of the issues here, making it the best representative anchor for this noisy cluster.", + "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43673", - "right": "issue:45341", + "left": "issue:27957", + "right": "issue:43352", "accept": false, - "reason": "Different bugs: missing GenerationMixin cache in chunked prefill vs a testing_utils issue." + "reason": "Both mention Flash Attention 2, but they are different model families and likely different support paths." }, { "left": "issue:43931", - "right": "issue:44410", + "right": "issue:44038", "accept": false, - "reason": "Both involve Qwen3-VL/Qwen3next, but one is a weight-shape loading mismatch and the other is missing layer projections; not the same defect." + "reason": "Qwen3-VL checkpoint shape mismatch is not the same bug as the broader Qwen3-VL-Moe transformers v5 failure." }, { - "left": "issue:42222", - "right": "issue:45542", + "left": "issue:43441", + "right": "issue:44829", "accept": false, - "reason": "VitPose model breakage and a TensorBoard-without-TensorFlow backend error are unrelated." + "reason": "One is an unsupported FlashAttention path; the other is degenerate training under flash_attention_3. Different concrete failures." }, { - "left": "issue:41628", - "right": "issue:45071", + "left": "issue:43012", + "right": "issue:43577", "accept": false, - "reason": "AutoImageProcessor import failure and PretrainedConfig type-checking regression are different code paths." + "reason": "A PyTorch precision warning on compile is unrelated to Blip2 dtype propagation staying float32." }, { - "left": "issue:45071", - "right": "issue:45542", + "left": "issue:43097", + "right": "issue:43452", "accept": false, - "reason": "Type checking in PretrainedConfig is unrelated to the TensorFlow backend import issue." + "reason": "Removing a public API is unrelated to gguf_file breaking tokenizer/model loading." }, { - "left": "issue:44484", - "right": "issue:45478", + "left": "issue:45003", + "right": "issue:45341", "accept": false, - "reason": "Unrelated issues: save_pretrained shard size question vs Qwen3.5-Moe from_pretrained error." + "reason": "sys.modules access in modeling_utils and a testing_utils bug are different code paths and symptoms." }, { - "left": "issue:27957", - "right": "issue:43352", + "left": "issue:43986", + "right": "issue:45341", "accept": false, - "reason": "Both mention Flash Attention 2, but one is XLM-RoBERTa support and the other is a model that explicitly does not support it; not the same bug." - }, - { - "left": "issue:43874", - "right": "issue:44261", - "accept": false, - "reason": "Different model-specific bugs: missing image-patch helper vs MLA RMS norm epsilon precision issue." - }, - { - "left": "issue:44263", - "right": "issue:46172", - "accept": false, - "reason": "Different DeepSeek/GLM internals and different failure modes." - }, - { - "left": "issue:44360", - "right": "issue:46172", - "accept": false, - "reason": "DSA indexer ReLU complaint is unrelated to the DeepSeek-V4 tensor transpose bug." - }, - { - "left": "issue:43931", - "right": "issue:44038", - "accept": false, - "reason": "Both are Qwen3-VL related, but the concrete problems differ: loading shape mismatch vs a broader version-specific failure." - }, - { - "left": "issue:43441", - "right": "issue:44829", - "accept": false, - "reason": "Both involve flash attention, but one is unsupported inference in Ministral-3 and the other is degenerate training with flash_attention_3." - }, - { - "left": "issue:43012", - "right": "issue:43577", - "accept": false, - "reason": "A compile-time bfloat16 warning is not the same bug as BLIP2 dtype remaining float32." - }, - { - "left": "issue:43097", - "right": "issue:43452", - "accept": false, - "reason": "Removal of a config helper and gguf_file loading breakage are unrelated." - }, - { - "left": "issue:45003", - "right": "issue:45341", - "accept": false, - "reason": "Unsafe sys.modules access and a testing_utils bug are different issues." - }, - { - "left": "issue:43986", - "right": "issue:45341", - "accept": false, - "reason": "AutoProcessor video loading crash is unrelated to testing_utils." + "reason": "Video processor crash without torchvision is unrelated to a testing utility issue." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "MLA layernorm epsilon precision issue and GLM-5 RoPE implementation are different model bugs." + "reason": "LayerNorm epsilon precision errors and RoPE implementation issues are distinct model internals." }, { "left": "issue:45753", "right": "issue:46369", "accept": false, - "reason": "Infinite-loop generation on one image is not the same as memory overflow in sam3-litetext." + "reason": "An infinite loop on a specific Qwen3_5 image is not the same as a sam3-litetext memory overflow." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "Tokenizer mapping None handling and physical-intelligence/fast loading failure are not the same underlying bug." + "reason": "Tokenizer mapping None in from_pretrained is a different loading bug than failing on physical-intelligence/fast." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and testing_utils bug are unrelated." + "reason": "A Qwen3-MoE sliding_window bug is unrelated to testing_utils." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Different GLM-related failures with different failing functions and symptoms." + "reason": "Missing image-patch count method and torch.split return handling are different GLM bugs." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "DeepseekV3MoE remote-vs-native divergence is broader and different from a specific DeepSeek-V4 tensor operation bug." + "reason": "DeepseekV3MoE divergence from remote impl is a different issue than the deepseek-v4 transpose/dtype bug." }, { - "left": "issue:43097", - "right": "issue:45588", + "left": "issue:43673", + "right": "issue:46123", "accept": false, - "reason": "Config removal and flash_attention.py crashing on s_aux=None are different defects." + "reason": "Generation cache missing in chunked_prefill is unrelated to the mask pipeline's final partial batch drop." }, { - "left": "issue:43673", - "right": "issue:46123", + "left": "issue:43097", + "right": "issue:45588", "accept": false, - "reason": "Generation cache missing during chunked prefill is unrelated to MaskGenerationPipeline dropping the final partial batch." + "reason": "A tie_embeddings API removal and a flash_attention.py s_aux=None crash are different problems." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Both are Qwen3-family issues, but one is missing attention projections and the other is a config field dropped by @strict." + "reason": "Missing Qwen3next projections and missing deepstack_visual_indexes are different model/config defects." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Missing torchvision crash and mistral_common ReasoningEffort import error are different dependency regressions." + "reason": "Missing torchvision processor crash is unrelated to the mistral_common ReasoningEffort import error." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Qwen3-VL-Moe failure and qwen3next missing projections are different problems." - }, - { - "left": "issue:43097", - "right": "issue:46050", - "accept": false, - "reason": "tie_embeddings removal and sinq save/load quantization breakage are unrelated." + "reason": "These affect different Qwen3 model variants with different missing pieces and failure modes." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they are for different symbols from different packages and likely different fixes." + "reason": "Both are import errors, but they concern different symbols and different dependency breakages." }, { "left": "issue:43208", "right": "issue:46897", "accept": false, - "reason": "xLSTM training bugs and Florence2 training-loss double-shift bug are unrelated model-specific defects." + "reason": "xLSTM training blockers and Florence2 training-loss double-shift are not the same underlying bug." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Wrong optimizer states under FSDP2/PEFT is a different training bug from Liger Kernel not being applied with model_init." + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to Liger Kernel not being applied in model_init." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "A processor crash without torchvision is unrelated to the save_pretrained shard-size question." }, { "left": "issue:46655", "right": "issue:46664", "accept": false, - "reason": "Both are SmolVLM video-processor issues, but padding calculation and num_frames semantics are different bugs." + "reason": "Both are SmolVLM video-processor issues, but one is padding calculation and the other is num_frames semantics." }, { "left": "issue:46298", "right": "issue:46752", "accept": false, - "reason": "Noisy logging and apply_chat_template crashing on an empty list are unrelated." + "reason": "A noisy logging message and an IndexError on empty chat-template input are not the same bug." + }, + { + "left": "issue:46172", + "right": "issue:46710", + "accept": false, + "reason": "The deepseek-v4 tensor-transpose issue is not the same as the later accuracy regression after upgrade." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "Missing set_seed in integration tests is unrelated to a noisy logging message." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "AddedToken keyword duplication and TOKENIZER_MAPPING_NAMES returning None are different tokenizer-loading failures." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "A removed tie_embeddings API and offloading failures under quantization are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and likely different compatibility breaks." + }, + { + "left": "issue:43012", + "right": "issue:46821", + "accept": false, + "reason": "A bfloat16 compile warning is unrelated to TimesFM window_size raising AttributeError." } ] }, @@ -16849,34 +16902,36 @@ "cluster_id": "cluster-10000-502", "nodes": [ "issue:36683", - "issue:38617", "issue:42175", - "issue:42890", + "issue:42490", + "issue:42757", "issue:42907", "issue:43012", + "issue:43064", "issue:43066", "issue:43097", "issue:43208", "issue:43352", "issue:43502", + "issue:43526", "issue:43550", "issue:43582", + "issue:43824", "issue:43825", - "issue:43873", "issue:43906", "issue:43976", "issue:43986", "issue:43992", "issue:44016", - "issue:44062", "issue:44077", "issue:44117", "issue:44162", "issue:44222", - "issue:44351", "issue:44368", "issue:44380", + "issue:44483", "issue:44484", + "issue:44514", "issue:44655", "issue:44704", "issue:44936", @@ -16893,17 +16948,9 @@ "issue:46298", "issue:46365", "issue:46421", - "issue:46701", - "issue:46710", - "issue:46821" + "issue:46701" ], "soft_pairs": [ - "issue:46172|issue:46710", - "issue:42890|issue:46298", - "issue:44062|issue:44117", - "issue:43097|issue:43873", - "issue:38617|issue:44351", - "issue:43012|issue:46821", "issue:44117|issue:46701", "issue:43502|issue:43992", "issue:44936|issue:46365", @@ -16918,246 +16965,245 @@ "issue:45712|issue:46298", "issue:43012|issue:44368", "issue:43208|issue:43550", - "issue:43352|issue:43976", "issue:43992|issue:44484", - "issue:43352|issue:44380", + "issue:43352|issue:43976", "issue:43066|issue:43906", + "issue:43352|issue:44380", "issue:44162|issue:46172", "issue:43097|issue:46298", "issue:45584|issue:46365", "issue:44380|issue:46132", "issue:44077|issue:46365", - "issue:36683|issue:43582", "issue:44222|issue:45478", + "issue:36683|issue:43582", "issue:44380|issue:45479", - "issue:42175|issue:46365" + "issue:42175|issue:46365", + "issue:42757|issue:43824", + "issue:43064|issue:44380", + "issue:42490|issue:43012", + "issue:44483|issue:44514", + "issue:43208|issue:43526" ], - "prepared_review_unit_hash": "af5f51156187654a82a7fbf1026deda301e41dc09182979d8fb7589f925a5de3", + "prepared_review_unit_hash": "9109958f0ecc265f3b47a4c02d0f0d25c87b974d81d5f460907284dfdd3514b4", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37413, - "estimated_input_tokens": 9354, - "estimated_eval_tokens": 18964 + "soft_pair_count": 32, + "serialized_chars": 37164, + "estimated_input_tokens": 9291, + "estimated_eval_tokens": 18838 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most items are distinct bug reports across unrelated subsystems, and none of the soft-similarity pairs look like safe duplicates. If a single issue must anchor the cluster, the quantization/offloading report is the most active and broadly representative, but only weakly so.", - "confidence": 0.27, - "canonical_issue_reason": "issue:43873 has the strongest discussion signal and inbound references, and it is a broad, actively reported runtime problem that can serve as the least-bad anchor for this noisy cluster.", + "summary": "This is not a duplicate cluster: the soft-similar pairs are mostly superficial keyword matches, but they point to different subsystems, APIs, or failure modes, so none should be merged.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue stands out; the items are unrelated bug reports rather than one underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is open, well-discussed, and thematically broad compared with the other items, so it is the most suitable representative issue in this set.", + "best_issue_reason": "No issue is a good global representative because the set spans many distinct code paths (loading, training, saving, tokenization, packaging, and model-specific failures).", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:46172", - "right": "issue:46710", - "accept": false, - "reason": "Both mention DeepSeek, but one is a tensor dtype/transpose bug and the other is an accuracy regression after a version upgrade; different code paths and failure modes." - }, - { - "left": "issue:42890", - "right": "issue:46298", - "accept": false, - "reason": "A flaky test due to missing seeding and a noisy logging message are unrelated." - }, - { - "left": "issue:44062", - "right": "issue:44117", - "accept": false, - "reason": "One is an AddedToken keyword conflict, the other is a tokenizer-mapping None assumption; different tokenizer bugs." - }, - { - "left": "issue:43097", - "right": "issue:43873", - "accept": false, - "reason": "A removed v5 config API and a quantization/offloading bug are not the same underlying issue." - }, - { - "left": "issue:38617", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but for different missing symbols and different parts of the package." - }, - { - "left": "issue:43012", - "right": "issue:46821", - "accept": false, - "reason": "A PyTorch warning on bfloat16 compilation and a TimesFM AttributeError are unrelated." - }, { "left": "issue:44117", "right": "issue:46701", "accept": false, - "reason": "Tokenizer mapping fallback logic and a pipeline failure for DiffusionGemmaGenerationOutput are different problems." + "reason": "Tokenizer mapping fallback/import handling is unrelated to the DiffusionGemma image-text pipeline failure." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "Local-files-only network leakage and a missing embed token weight during UMT5 loading are distinct loader bugs." + "reason": "Local-files-only network leakage and a missing embed_tokens weight during from_pretrained are different loading bugs." }, { "left": "issue:44936", "right": "issue:46365", "accept": false, - "reason": "Trainer evaluate-after-train failure and an import conflict with datasets are different failure modes." + "reason": "train/evaluate lifecycle failure and importing Trainer being blocked by datasets are separate problems." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "Passing kwargs to cached_file and an auto-docstring crash from future annotations are unrelated." + "reason": "cached_file kwargs forwarding and docstring processing with future annotations touch different code paths." }, { "left": "issue:44380", "right": "issue:46421", "accept": false, - "reason": "Both involve backend/runtime behavior, but one ignores attention scaling while the other causes recompilation from cache lazy init." + "reason": "GPT2 attention scaling under SDPA/FlashAttention is unrelated to chunked-prefill recompilation from cache lazy init." }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 import/load failure and an invalid tensorflow_text backend mapping are unrelated." + "reason": "Python 3.14 import/load failure and tensorflow_text backend validation are not the same defect." }, { "left": "issue:43986", "right": "issue:44016", "accept": false, - "reason": "A missing torchvision dependency path and a notebook syntax error are not duplicates." + "reason": "AutoProcessor without torchvision and a notebook syntax error are completely unrelated." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals and an infinite loop on a specific Qwen3_5 image are different bugs." + "reason": "Saving dequantized Ministrals/Devstrals and a Qwen3.5 image loop are different model-specific issues." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "A model-save bug and a pipeline serialization bug both touch saving, but they are different object types and code paths." + "reason": "Saving dequantized weights and saving Pipeline objects are both save-related but not the same underlying bug." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "An error-message wording issue for pipeline translation support is unrelated to pipeline save_pretrained behavior." + "reason": "Bad pipeline() task messaging and save_pretrained for Pipeline objects are unrelated API issues." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "A warning about bfloat16 precision and a Bamba torch.compile SDPA failure are not the same bug." + "reason": "A bfloat16 compilation warning is not the same as a Bamba torch.compile SDPA failure." }, { "left": "issue:45712", "right": "issue:46298", "accept": false, - "reason": "Leftover dummy classes and a noisy logging message are unrelated repository hygiene issues." + "reason": "Leftover dummy classes and a noisy logging message are unrelated maintenance issues." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "Different warnings affecting different config fields; not the same underlying defect." + "reason": "Compilation precision warnings and tie_word_embeddings warning spam come from different model behaviors." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training bugs for small models and a Bamba compile/SDPA failure are unrelated." - }, - { - "left": "issue:43352", - "right": "issue:43976", - "accept": false, - "reason": "A model-specific Flash Attention limitation and a Python-version compatibility issue are unrelated." + "reason": "xLSTM small-model training bugs do not match the Bamba SDPA torch.compile failure." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing weights on load and a question about default shard size are not duplicates." + "reason": "Missing checkpoint weights and the 50GB max_shard_size default concern different save/load behavior." }, { "left": "issue:43352", - "right": "issue:44380", + "right": "issue:43976", "accept": false, - "reason": "Both involve attention/backends, but one is a model support limitation and the other is backend behavior ignoring a config parameter." + "reason": "Flash Attention 2.0 unsupported for a model and a Python version compatibility problem are unrelated." }, { "left": "issue:43066", "right": "issue:43906", "accept": false, - "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue do not clearly match the same bug." + "reason": "Tokenizer decoder type handling and an isolated reproduction of another issue are not the same bug." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Unsupported FA2 on a model and GPT2 attention scaling being ignored under backends are different backend/model issues." }, { "left": "issue:44162", "right": "issue:46172", "accept": false, - "reason": "ESM2 breakage and a DeepSeek tensor operation bug are unrelated." + "reason": "ESM2 breakage and a deepseek transpose/dtype bug are separate model-specific failures." }, { "left": "issue:43097", "right": "issue:46298", "accept": false, - "reason": "A removed API and a noisy log message are unrelated." + "reason": "Removal of a public config flag and a noisy logging message are unrelated." }, { "left": "issue:45584", "right": "issue:46365", "accept": false, - "reason": "Whisper empty-transcription generation failure and a Trainer import conflict are unrelated." + "reason": "Whisper empty-transcription generation failure and importing Trainer affected by datasets are unrelated." }, { "left": "issue:44380", "right": "issue:46132", "accept": false, - "reason": "Attention scaling being ignored and AttentionInterface.register changing registered-function behavior are different attention-stack issues." + "reason": "GPT2 SDPA/FlashAttention scaling and AttentionInterface.register behavior changes are different attention-related issues." }, { "left": "issue:44077", "right": "issue:46365", "accept": false, - "reason": "PatchTSMixer post_init validation and a Trainer import conflict are unrelated." + "reason": "patchtsmixer post_init policy and Trainer import breakage have no code-path overlap." }, { - "left": "issue:36683", - "right": "issue:43582", + "left": "issue:44222", + "right": "issue:45478", "accept": false, - "reason": "A missing config attribute and an Apple Silicon TypeError in allocator warmup are unrelated." + "reason": "FP8 save_pretrained for MoE and Qwen3.5 MoE from_pretrained errors are related by model family but not the same concrete failure." }, { - "left": "issue:44222", - "right": "issue:45478", + "left": "issue:36683", + "right": "issue:43582", "accept": false, - "reason": "Both mention MoE/Qwen, but one is about FP8 save_pretrained and the other is a from_pretrained error; different code paths." + "reason": "Gemma3Config missing vocab_size and an Apple Silicon TypeError in caching_allocator_warmup are unrelated." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Backend attention scaling and a zero-loss classification bug are unrelated." + "reason": "Attention scaling under SDPA/FlashAttention is not the same as a degenerate zero-loss classification bug." }, { "left": "issue:42175", "right": "issue:46365", "accept": false, - "reason": "Missing TensorFlow in a torch extra install and a Trainer import conflict after uninstalling datasets are unrelated dependency issues." + "reason": "TensorFlow missing from a torch extra install and Trainer import failing because of datasets are unrelated packaging/import issues." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Offline-mode import failure from huggingface_hub and Qwen2_5_VL symbol import failure are separate missing-symbol problems." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer state corruption and GPT2 attention scaling under backend dispatch are different training vs inference bugs." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local path trailing slash warning and bfloat16 compilation precision warning are unrelated loading vs compile diagnostics." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat/completions request rejection and batched Qwen2_5_VLProcessor template crashes are different request paths." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training bugs and BeitImageProcessorFast reduce_labels truncation are unrelated model vs image-processor issues." } ] }, @@ -17186,11 +17232,9 @@ "issue:36296", "issue:36683", "issue:38617", - "issue:42490", "issue:42757", "issue:42994", "issue:43010", - "issue:43012", "issue:43064", "issue:43208", "issue:43299", @@ -17198,16 +17242,16 @@ "issue:43475", "issue:43502", "issue:43525", - "issue:43526", "issue:43611", "issue:43618", "issue:43756", "issue:43761", - "issue:43824", "issue:43867", "issue:43937", "issue:43976", + "issue:43992", "issue:44038", + "issue:44075", "issue:44077", "issue:44230", "issue:44355", @@ -17215,28 +17259,25 @@ "issue:44483", "issue:44485", "issue:44492", - "issue:44514", "issue:44625", "issue:44655", + "issue:44908", "issue:44938", "issue:45216", "issue:45276", "issue:45278", + "issue:45310", "issue:46050", "issue:46077", "issue:46082", "issue:46365", "issue:46369", "issue:46829", + "issue:46851", "issue:46877" ], "soft_pairs": [ - "issue:42757|issue:43824", - "issue:43064|issue:44380", - "issue:42490|issue:43012", - "issue:44483|issue:44514", "issue:43618|issue:44380", - "issue:43208|issue:43526", "issue:38617|issue:43502", "issue:43756|issue:44485", "issue:43618|issue:43761", @@ -17246,8 +17287,8 @@ "issue:36010|issue:42757", "issue:44625|issue:46365", "issue:45276|issue:46077", - "issue:44483|issue:46369", "issue:43611|issue:44655", + "issue:44483|issue:46369", "issue:43352|issue:43937", "issue:36683|issue:43475", "issue:44355|issue:46877", @@ -17263,244 +17304,232 @@ "issue:36683|issue:44625", "issue:29870|issue:36296", "issue:43010|issue:45278", - "issue:43064|issue:44492" + "issue:43064|issue:44492", + "issue:43992|issue:45310", + "issue:44075|issue:46365", + "issue:44908|issue:46851" ], - "prepared_review_unit_hash": "e37a7c797ff02933efcd1997fa3faa02326cd30db4b822d8ab7f6bd1320f3c3a", + "prepared_review_unit_hash": "52a9f4a0c4adf7aaa807fb549d6b5622c8786456543306b09bcc448080d2a37e", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37304, - "estimated_input_tokens": 9326, - "estimated_eval_tokens": 18908 + "soft_pair_count": 31, + "serialized_chars": 36574, + "estimated_input_tokens": 9144, + "estimated_eval_tokens": 18544 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue reports. The only likely duplicate soft edge is the CLIP v5 regression around missing vision-model outputs; the rest differ in model, API surface, or failure mode.", - "confidence": 0.71, - "canonical_issue_reason": "issue:43761 is the clearest representative of the CLIP regression: it names the exact code path (`CLIPVisionModel.forward`) and the missing output behavior under `output_hidden_states=True`.", + "summary": "This cluster is a grab bag of unrelated issues. The soft pairs mostly share broad keywords like import error, config, save/load, or model family, but they do not describe the same underlying bug or fix.", + "confidence": 0.88, + "canonical_issue_reason": "No single issue is a good canonical representative here: the items span many distinct model families and failure modes, so treating any one as the duplicate target would be misleading.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43761 is the best single issue to anchor this small CLIP duplicate cluster because it is specific, reproducible, and points to the concrete forward-path regression.", + "best_issue_reason": "No global best issue candidate stands out because none of the soft pairs are true duplicates; the safest triage is to keep them separate.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42757", - "right": "issue:43824", - "accept": false, - "reason": "Both are import errors, but for different symbols and packages (`huggingface_hub` vs `transformers`); not the same bug." - }, - { - "left": "issue:43064", - "right": "issue:44380", - "accept": false, - "reason": "Different failures: distributed optimizer state corruption vs attention scaling being ignored in specific backends." - }, - { - "left": "issue:42490", - "right": "issue:43012", - "accept": false, - "reason": "Both mention warnings, but one is a path-loading import warning and the other is a PyTorch precision warning during compile." - }, - { - "left": "issue:44483", - "right": "issue:44514", - "accept": false, - "reason": "One is an API request validation problem; the other is a chat-template crash on batched input." - }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIP output regression vs GPT-2 attention scaling behavior; different model families and code paths." - }, - { - "left": "issue:43208", - "right": "issue:43526", - "accept": false, - "reason": "Unrelated subsystems: xLSTM training bugs vs BeitImageProcessorFast label reduction." + "reason": "Both are attention-related regressions, but one is CLIP attentions missing and the other is GPT2 attention scaling ignored under SDPA/FlashAttention; different code paths." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "Different import/load paths: missing `layer_type_validation` export vs local-files-only network requests." + "reason": "An import error for configuration_utils is unrelated to local_files_only still triggering network requests." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Both mention RoPE, but one is a Smollm3 layer-selection bug and the other is a GLM-5 RoPE implementation discussion." + "reason": "Smollm3 RoPE-layer dropping and GLM-5 RoPE implementation discussion are different model-specific problems." }, { "left": "issue:43618", "right": "issue:43761", - "accept": true, - "reason": "Both describe the same CLIP v5 regression in output assembly: missing vision-model outputs/fields in forward pass." + "accept": false, + "reason": "Both involve CLIP, but one is missing attentions and the other is hidden_states not returned; related area, not the same bug." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Different save_pretrained failures affecting different object types (quantized models vs Pipeline objects)." + "reason": "Both involve saving, but quantized model save failure and Pipeline save_pretrained are different serialization paths." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "Same model family, but one is a v5/Qwen3-VL-Moe bug report and the other is fp8 inference support; not the same concrete issue." + "reason": "Qwen3-VL-Moe loading regression and fp8 support for Qwen3-VL/Qwen3.5 MoE are different changes and failure modes." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "One is a Qwen3VL-Moe loading regression, the other is a load failure when state_dict keys are sorted; insufficiently specific and likely different causes." + "reason": "Qwen3-VL-Moe loading breakage and sorted state_dict load errors are separate checkpoint-loading issues." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they concern different removed symbols from different dependencies." + "reason": "Both are import errors, but GenerationMixin and is_offline_mode are unrelated symbols from different modules." }, { "left": "issue:44625", "right": "issue:46365", "accept": false, - "reason": "Qwen3.5 config propagation bug vs inability to import Trainer due to a datasets conflict." + "reason": "Qwen3.5 num_labels propagation and failing to import Trainer because of datasets are unrelated." }, { "left": "issue:45276", "right": "issue:46077", "accept": false, - "reason": "Gemma4 embedding resize bug vs Gemma4 attention-mask construction bug; different behaviors and fixes." - }, - { - "left": "issue:44483", - "right": "issue:46369", - "accept": false, - "reason": "Chat-completions request rejection and SAM memory overflow are unrelated." + "reason": "Gemma4 resize_token_embeddings behavior and bidirectional-attention mask construction are different Gemma4 bugs." }, { "left": "issue:43611", "right": "issue:44655", "accept": false, - "reason": "Model-loading/base_model_prefix regression vs Pipeline save_pretrained failure; different code paths." + "reason": "Broken loading with base_model_prefix and Pipeline save_pretrained are separate model-loading vs pipeline-serialization issues." + }, + { + "left": "issue:44483", + "right": "issue:46369", + "accept": false, + "reason": "A chat-completions API request validation issue is unrelated to SAM3 model memory overflow." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "Unsupported Flash Attention for Nemotron vs invalid GenerationConfig for GLM-5; different errors." + "reason": "NemotronH FlashAttention2 support and GLM-5 GenerationConfig validation are different model/config problems." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Different attribute-missing bugs in different models (Gemma3Config vs SAM3 vision output)." + "reason": "Gemma3Config missing vocab_size and SAM3VisionEncoderOutput missing fpn_position_embeddings are unrelated attribute bugs." }, { "left": "issue:44355", "right": "issue:46877", "accept": false, - "reason": "Running compiled Python files failing is unrelated to auto_docstring printing diagnostics at import time." + "reason": "Compiled Python file errors and auto_docstring emitting doc-lint output at import time are not the same issue." }, { "left": "issue:44655", "right": "issue:46050", "accept": false, - "reason": "Both involve saving/loading, but one is pipeline serialization and the other is quantized-model roundtrip corruption." + "reason": "Pipeline save_pretrained and quantized sinq round-trip breakage both touch persistence, but they are distinct code paths and bugs." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Gemma3 config attribute bug and `patchtsmixer` post_init validation are unrelated." + "reason": "Missing vocab_size on Gemma3Config and patchtsmixer post_init validation are different configuration issues." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Both are version/platform compatibility complaints, but for different Python versions and failure modes." + "reason": "These are both version-compatibility reports, but Python 3.9+ vs Python 3.14 failures are unrelated." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Config propagation for Qwen3.5 is unrelated to Gemma4 attention-mask behavior." + "reason": "num_labels propagation and Gemma4 bidirectional attention masking are different configuration behaviors." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "A sorted-state_dict load issue and a Qwen3.5 save_pretrained regression are different problems." + "reason": "State_dict ordering causing load errors and Qwen3.5 save_pretrained corruption are different checkpoint bugs." }, { "left": "issue:46077", "right": "issue:46082", "accept": false, - "reason": "One is bidirectional-attention mask construction, the other is Llama head_dim validation." + "reason": "Gemma4 attention-mask behavior and LlamaConfig head_dim validation are unrelated config issues." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP output regression vs Qwen3.5 checkpoint saving corruption; unrelated." + "reason": "CLIP attention output regression and Qwen3.5 saved-checkpoint regression are unrelated models and symptoms." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Different missing-config-attribute bugs in different model configs." + "reason": "Gemma3Config missing vocab_size and Llama4Config missing pad_token_id are separate model config attribute problems." }, { "left": "issue:43208", "right": "issue:46829", "accept": false, - "reason": "xLSTM training blockers vs VideoPrism output field mix-up are unrelated." + "reason": "xLSTM training bugs and VideoPrism hidden_states mapping are unrelated model-specific issues." }, { "left": "issue:36683", "right": "issue:44625", "accept": false, - "reason": "Both are config propagation/attribute issues, but for different models and fields." + "reason": "Both are config attribute propagation bugs, but Gemma3 vocab_size and Qwen3.5 num_labels are distinct fields in different models." }, { "left": "issue:29870", "right": "issue:36296", "accept": false, - "reason": "Bart generation UnboundLocalError and tensor-parallel training bug are different subsystems and failure modes." + "reason": "BART generate() decoder_attention_mask UnboundLocalError and tensor-parallel training bug are unrelated failures." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Cache/no_grad decorator issue and broad import-error report after upgrade are not the same bug." + "reason": "Cache/update no_grad decoration and broad import errors after upgrade do not describe the same bug." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "Distributed optimizer state bug vs cache-strategy typo; no underlying overlap." + "reason": "FSDP2/PEFT optimizer-state corruption and a typo in cache strategies are unrelated." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "UMT5Encoder missing embed_tokens.weight and Qwen3.5 MoE from_pretrained error are different model-loading bugs." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Ignored SGD args and inability to import Trainer due to datasets are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:46851", + "accept": false, + "reason": "Scheduler kwargs being ignored and Qwen3.5 variable-length position_ids are separate issues." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet, and the soft-edge decisions are generally conservative. The one accepted CLIP pair is plausibly the same forward-output regression family, while the rest are appropriately rejected as different bugs, APIs, or failure modes." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -17529,29 +17558,30 @@ "issue:43064", "issue:43502", "issue:43575", + "issue:43576", "issue:43618", "issue:43723", "issue:43824", "issue:43867", + "issue:43874", "issue:43976", "issue:43992", "issue:44062", - "issue:44075", "issue:44077", "issue:44112", "issue:44222", "issue:44246", "issue:44355", "issue:44360", + "issue:44387", + "issue:44485", "issue:44556", "issue:44734", "issue:44855", - "issue:44908", "issue:44938", "issue:44964", "issue:45071", "issue:45125", - "issue:45310", "issue:45357", "issue:45375", "issue:45440", @@ -17559,17 +17589,13 @@ "issue:45561", "issue:45600", "issue:45850", + "issue:46132", "issue:46202", "issue:46285", - "issue:46365", "issue:46617", - "issue:46664", - "issue:46851" + "issue:46664" ], "soft_pairs": [ - "issue:43992|issue:45310", - "issue:44075|issue:46365", - "issue:44908|issue:46851", "issue:29870|issue:33453", "issue:44938|issue:45071", "issue:44355|issue:44855", @@ -17594,211 +17620,221 @@ "issue:43502|issue:44062", "issue:44734|issue:46617", "issue:43723|issue:44556", + "issue:43824|issue:46617", "issue:44222|issue:45125", - "issue:43824|issue:46617" + "issue:44387|issue:44556", + "issue:43502|issue:44485", + "issue:43874|issue:46617", + "issue:43576|issue:46132" ], - "prepared_review_unit_hash": "2d16ee41972670553f1fd024548f0369a5326a8ea7df7ffb9442b28708358566", + "prepared_review_unit_hash": "8ec33564fd2af8b73bebc22b0f4e8f23004c44744464027d120f21018279995b", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36078, - "estimated_input_tokens": 9020, - "estimated_eval_tokens": 18296 + "soft_pair_count": 30, + "serialized_chars": 36440, + "estimated_input_tokens": 9110, + "estimated_eval_tokens": 18476 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers bug reports centered on loading/import/version-regression issues. None of the soft pairs look like true duplicates, so I reject all of them.", - "confidence": 0.77, - "canonical_issue_reason": "issue:33453 is the broadest representative of the cluster\u2019s recurring loading-regression theme (tokenizer/model reload problems under newer versions).", + "summary": "Mostly a heterogeneous issue set; the soft pairs share broad themes (loading regressions, import-time failures, tensor-parallel bugs, CI flakiness) but do not look like the same concrete underlying bug/change.", + "confidence": 0.79, + "canonical_issue_reason": "issue:33453 is the most generic, cluster-like representative here: a broad tokenizer-loading regression title that could plausibly sit at the center of the loading-related subset.", "canonical_pr_reason": null, - "best_issue_reason": "issue:33453 is the most generic and representative issue title here, making it the best anchor for a heterogeneous loading-regression cluster.", + "best_issue_reason": "issue:33453 is the best overall issue representative because it is broad and concise, making it a reasonable anchor for the few loading-related items in this otherwise mixed cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43992", - "right": "issue:45310", - "accept": false, - "reason": "Different models and different missing-field/load failures (UMT5 encoder weight vs Qwen3.5 MoE from_pretrained)." - }, - { - "left": "issue:44075", - "right": "issue:46365", - "accept": false, - "reason": "Optimizer argument handling is unrelated to a Trainer import failure involving datasets." - }, - { - "left": "issue:44908", - "right": "issue:46851", - "accept": false, - "reason": "Scheduler kwargs bug is unrelated to Qwen3.5 variable-length position_ids behavior." - }, { "left": "issue:29870", "right": "issue:33453", "accept": false, - "reason": "Bart generate decoder_attention_mask bug is a different code path from tokenizer-loading regression." + "reason": "Different failures: BART generate() decoder_attention_mask bug vs tokenizer loading regression." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Both are version-compatibility complaints, but one is Python 3.14 import failure and the other is PretrainedConfig type checking; not the same bug." + "reason": "Unrelated problems: Python 3.14 import/load failure vs PretrainedConfig type-checking regression." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Compiled-Python-file errors are unrelated to the DebertaV2 Python 3.13 torch.jit.script parsing issue." + "reason": "Different import-time issues with different triggers and code paths." }, { "left": "issue:43867", "right": "issue:46202", "accept": false, - "reason": "State-dict sorting load error and Gemma pipeline roundtrip failure are different problems." + "reason": "State_dict sorting load error is not the same as Gemma pipeline roundtrip token issue." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "Gemma3Config missing vocab_size is unrelated to the FSDP2/PEFT optimizer-state bug." + "reason": "Gemma3Config missing vocab_size vs FSDP2/PEFT optimizer-state corruption are unrelated." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Both are test-related, but one is seed flakiness and the other is an xdist file race; different root causes." + "reason": "Both are flaky-test themed, but one is missing set_seed and the other is an xdist race on captured_info.txt." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "State-dict sorting load error does not match the missing vision-config field issue." + "reason": "Different models and different missing fields/serialization bugs." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "No_grad decoration on cache/layer updates is unrelated to the broad 'breaks too many models' report." + "reason": "no_grad decoration fix vs vague broad breakage report are not the same bug." }, { "left": "issue:41669", "right": "issue:44246", "accept": false, - "reason": "Import-star cleanup/performance work is not the same as intermittent slow import reports." + "reason": "Both mention import time, but one is import-star cleanup and the other is intermittent slowdown; too broad to merge." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bug and TP-related OOM on model load are distinct failures." + "reason": "Tensor-parallel training bug and tp OOM on loading a specific model are different concrete failures." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "Cache update decoration issue is unrelated to the Python version compatibility problem." + "reason": "Different issue types: no_grad update methods vs Python/version compatibility regression." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "CLIP attentions assignment and patchtsmixer post_init validation are unrelated model-specific issues." + "reason": "Different model API regressions; no shared code path is evident." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "State-dict ordering load failure is unrelated to incorrect visual encoder keys in save_pretrained." + "reason": "Load-state_dict error and incorrect save_pretrained visual keys are different serialization paths." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "These are separate Qwen/encoder loading errors with different missing data and models." + "reason": "UMT5Encoder missing embed_tokens.weight vs Qwen3.5 MoE from_pretrained error are different model-specific load failures." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "Missing PyTorch dependency import error is not a tokenizer-loading regression." + "reason": "Missing PyTorch import dependency is unrelated to tokenizer-loading regression." }, { "left": "issue:43618", "right": "issue:46285", "accept": false, - "reason": "Both mention CLIP, but one is missing attentions assignment and the other is class-structure drift; not the same bug." + "reason": "CLIP attentions assignment vs CLIP text-model structure mismatch are separate regressions." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "DSA indexer ReLU discussion is unrelated to DeepseekV3MoE implementation divergence." + "reason": "DSA indexer activation bug and DeepseekV3MoE implementation divergence are unrelated." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Seed-related test flakiness is unrelated to loading Phi-4 multimodal model." + "reason": "Test seeding flakiness vs Phi-4-multimodal model load failure are not the same bug." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Test seed flakiness and GraniteSpeech stale device override are different CI issues." + "reason": "Different CI/test failures with different causes." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "Seed-setting test failures are unrelated to stale auto_mappings references." + "reason": "Different CI regressions: seedless flaky integration test vs stale auto_mappings reference." }, { "left": "issue:42898", "right": "issue:46664", "accept": false, - "reason": "Tokenization-space behavior in v5 is unrelated to SmolVLM num_frames semantics." + "reason": "Tokenization-space cleanup behavior and video num_frames semantics are unrelated." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Offline loading making API requests is unrelated to AddedToken special-argument duplication." + "reason": "Offline local_files_only API leakage is unrelated to AddedToken keyword argument collision." }, { "left": "issue:44734", "right": "issue:46617", "accept": false, - "reason": "KV-cache continuation crash in serve is unrelated to trust_remote_code cache FileNotFoundError." + "reason": "KV-cache continuation indexing crash vs cache file not found are different code paths." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both are load-related regressions, but one is tokenizer loading and the other is checkpoint reload after upgrade; not the same concrete bug." + "reason": "Tokenizer loading regression and model checkpoint reload incompatibility are related only at a very high level, not the same bug." + }, + { + "left": "issue:43824", + "right": "issue:46617", + "accept": false, + "reason": "Missing import symbol and trust_remote_code cache FileNotFoundError are different issues." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "FP8 save_pretrained bug is unrelated to missing tensor-parallel plan metadata." + "reason": "FP8 save_pretrained bug and missing tensor-parallel plan are unrelated model features." }, { - "left": "issue:43824", + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization memory regression and v4-to-v5 reload failure are different failures." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only network leakage has nothing to do with GLM-5 RoPE implementation." + }, + { + "left": "issue:43874", "right": "issue:46617", "accept": false, - "reason": "Missing Qwen2_5_VL import and trust_remote_code cache FileNotFoundError are different failure modes." + "reason": "Missing image-patch helper and trust_remote_code cache lookup failure are different multimodal loading issues." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "transformers env CLI breakage and AttentionInterface.register behavior change are separate features." } ] }, @@ -17840,9 +17876,9 @@ "issue:43475", "issue:43502", "issue:43576", + "issue:43673", "issue:43824", "issue:43827", - "issue:43874", "issue:43901", "issue:44222", "issue:44230", @@ -17850,8 +17886,8 @@ "issue:44351", "issue:44355", "issue:44360", - "issue:44387", "issue:44485", + "issue:44492", "issue:44556", "issue:44749", "issue:44908", @@ -17859,10 +17895,11 @@ "issue:44961", "issue:45071", "issue:45125", + "issue:45362", "issue:45375", "issue:45468", "issue:45478", - "issue:46132", + "issue:45850", "issue:46143", "issue:46144", "issue:46209", @@ -17871,22 +17908,18 @@ "issue:46726" ], "soft_pairs": [ - "issue:44387|issue:44556", - "issue:43502|issue:44485", - "issue:43874|issue:46617", - "issue:43576|issue:46132", "issue:43010|issue:43065", "issue:30333|issue:33453", "issue:45125|issue:45478", "issue:44355|issue:44749", "issue:45125|issue:45375", "issue:30990|issue:31795", - "issue:46664|issue:46726", "issue:43010|issue:43232", + "issue:46664|issue:46726", "issue:43502|issue:44961", "issue:43502|issue:44222", - "issue:43827|issue:44961", "issue:42548|issue:44351", + "issue:43827|issue:44961", "issue:30333|issue:36296", "issue:43299|issue:44556", "issue:29870|issue:31795", @@ -17908,271 +17941,264 @@ "issue:44315|issue:46209", "issue:44222|issue:44230", "issue:41628|issue:42548", - "issue:44485|issue:46664" + "issue:44485|issue:46664", + "issue:43673|issue:44492", + "issue:45362|issue:45850" ], - "prepared_review_unit_hash": "65c17b629c68a2a1a31d9f5f28fc5ba1c024e37d8b1b9f7dd55b556668343960", + "prepared_review_unit_hash": "39db6bd8a87d7dd74f1393e6fb6b24c2785293f1d2401798f0b098dd3d1e39b1", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 38, - "serialized_chars": 37993, - "estimated_input_tokens": 9499, - "estimated_eval_tokens": 19254 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 37785, + "estimated_input_tokens": 9447, + "estimated_eval_tokens": 19150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: it mixes unrelated bug reports, import errors, docs issues, and model-specific regressions. I found no soft pair that is clearly the same underlying issue/change, so all candidate duplicates are rejected.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43576 is the closest thing to a center only because it appears in multiple weak similarity links, but it is still a specific v5 CLI regression and not a strong canonical for the whole cluster.", + "summary": "Mostly a heterogeneous cluster of unrelated Transformers issues. Only a couple of soft links look plausibly duplicate: Qwen3.5-Moe loading regressions and a pair of RoFormer attention/kwargs issues. Most other pairs are clearly different bugs, commands, docs, or model families.", + "confidence": 0.41, + "canonical_issue_reason": "issue:45125 is the clearest concrete bug report in the one strong duplicate pair here: it names the likely root cause (`_tp_plan` missing) behind the Qwen3.5-Moe loading failure.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43576 is the least-bad representative among the set due to repeated soft matches, but the cluster is too mixed for it to be a true duplicate canonical.", + "best_issue_reason": "issue:45125 is the best single anchor among the items because it is specific, reproducible, and the strongest candidate for grouping with a closely related loading error report.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44387", - "right": "issue:44556", - "accept": false, - "reason": "Both mention v5 breakage, but one is an int4 quantization memory/OOM regression and the other is a checkpoint reload incompatibility; different code paths." - }, - { - "left": "issue:43502", - "right": "issue:44485", - "accept": false, - "reason": "Local-files-only API calls and GLM RoPE implementation are unrelated bugs." - }, - { - "left": "issue:43874", - "right": "issue:46617", - "accept": false, - "reason": "One is a missing image-processor method causing token accounting failure; the other is a cache FileNotFoundError for trust_remote_code loading." - }, - { - "left": "issue:43576", - "right": "issue:46132", - "accept": false, - "reason": "A broken env command and an AttentionInterface.register behavior change are different features and failure modes." - }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "Cache/layer no_grad decoration and Sam3PixelDecoder's dummy Conv2d are unrelated changes." + "reason": "Different subsystems: generation cache update/no-grad vs Sam3PixelDecoder dummy Conv2d." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLflow reporting on failed jobs is unrelated to tokenizer loading regressions." + "reason": "MLflow job-reporting failure and tokenizer-loading regression are unrelated." }, { "left": "issue:45125", "right": "issue:45478", - "accept": false, - "reason": "Both are Qwen3.5-Moe-related, but tensor-parallel plan omission and from_pretrained loading failure are distinct issues." + "accept": true, + "reason": "Both describe the same Qwen3.5-Moe loading path failing; 45125 gives the likely root cause and 45478 is a broader symptom report that could be fixed together." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled Python file errors and data filtering slowdown after upgrade are not the same bug." + "reason": "Compiled-file execution errors and a data-filtering slowdown are unrelated." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "Missing _tp_plan and a missing vision-config field are different model/config problems." + "reason": "One is tensor-parallel model-loading support; the other is a missing config field in Qwen3.5-Moe-Vision." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Sentence-transformers loading hang and forward() doc confusion are unrelated." + "reason": "Sentence-Transformers loading hang and forward() documentation confusion are not the same bug." }, { - "left": "issue:46664", - "right": "issue:46726", + "left": "issue:43010", + "right": "issue:43232", "accept": false, - "reason": "A video num_frames semantic inconsistency and an AutoProcessor backend-missing error are different problems." + "reason": "Both are generation-related, but one is about cache update decoration and the other about sync_gpus model-kwargs handling." }, { - "left": "issue:43010", - "right": "issue:43232", + "left": "issue:46664", + "right": "issue:46726", "accept": false, - "reason": "Both touch generation/cache code, but one is about no_grad on update() and the other about sync_gpus handling in _update_model_kwargs_for_generation." + "reason": "Video processor frame-count semantics and missing vision-backend import behavior are different issues." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "Local-files-only network behavior and an unrelated 'racoon' issue do not match." + "reason": "Local-files-only network leak and an unrelated 'racoon' issue do not match." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "API requests despite local_files_only and FP8 save_pretrained are different bugs." + "reason": "Offline-loading bug and FP8 save_pretrained/moe behavior are different code paths." }, { - "left": "issue:43827", - "right": "issue:44961", + "left": "issue:42548", + "right": "issue:44351", "accept": false, - "reason": "Docs still referencing pipeline() and 'racoon' are unrelated." + "reason": "Both are import errors, but for different exported symbols with no clear shared root cause." }, { - "left": "issue:42548", - "right": "issue:44351", + "left": "issue:43827", + "right": "issue:44961", "accept": false, - "reason": "Different import errors for different symbols; no evidence of a single underlying export breakage." + "reason": "Docs still mentioning pipeline() and the unrelated 'racoon' report are not duplicates." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "Failed-job MLflow reporting and tensor-parallel training are unrelated." + "reason": "MLflow failure reporting and tensor-parallel training bug are unrelated." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Qwen3VL-Moe loading regression and generic v4.57-to-v5 reload failure are not the same concrete bug." + "reason": "Both are version-regression loading problems, but they involve different models and failure modes." }, { "left": "issue:29870", "right": "issue:31795", "accept": false, - "reason": "A generate() UnboundLocalError and forward() documentation confusion are unrelated." + "reason": "An UnboundLocalError in generate() is unrelated to forward() documentation confusion." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-file execution errors and PretrainedConfig type checking are different issues." + "reason": "Compiled-file execution errors and PretrainedConfig type-checking regressions are different bugs." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Tokenizer loading regression and Llama 3.2 Vision model loading regression are separate loading failures." + "reason": "Tokenizer loading regression and Llama 3.2 vision model-loading breakage are different symptoms." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Broken env command and PretrainedConfig type checking are not the same bug." + "reason": "Broken `transformers env` command and PretrainedConfig type checking are unrelated." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Missing huggingface_hub symbol import and broken transformers env CLI are unrelated." + "reason": "Missing `is_offline_mode` import from huggingface_hub is not the same as the v5 env-command breakage." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger kernel application in model_init and a ReLU omission in DSA indexer are different code paths." + "reason": "Liger-kernel application in `model_init` and a DSA indexer/ReLU discussion do not match." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion and an ImportError for layer_type_validation are unrelated." + "reason": "Docs confusion in model.forward and an ImportError for `layer_type_validation` are unrelated." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Broken env command and missing Qwen2.5-VL export are unrelated import/CLI problems." + "reason": "Broken env command and Qwen2.5-VL import failure are different issues." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "No meaningful overlap between 'racoon' and Gemma-4 audio positional encoding." + "reason": "The placeholder 'racoon' issue and Gemma4AudioRelPositionalEncoding bug are unrelated." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation and 'racoon' are unrelated." + "reason": "GLM-5 RoPE implementation and the unrelated 'racoon' issue do not overlap." }, { "left": "issue:44351", "right": "issue:46617", "accept": false, - "reason": "HybridCache import failure and trust_remote_code cache FileNotFoundError are different failure modes." + "reason": "Top-level import failure for HybridCache and a cache FileNotFoundError are distinct failure modes." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor-parallel training bug and text-classification docs mismatch are unrelated." + "reason": "Tensor-parallel training and pipeline doc wording are unrelated." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs handling and 'racoon' are unrelated." + "reason": "Scheduler kwargs handling and the unrelated 'racoon' issue are not duplicates." }, { "left": "issue:43576", "right": "issue:44938", "accept": false, - "reason": "Broken env command and Python 3.14 load failure are different issues." + "reason": "Broken env command and Python 3.14 loading failure are different regressions." }, { "left": "issue:43475", "right": "issue:46617", "accept": false, - "reason": "SAM 3 video attribute error and trust_remote_code cache FileNotFoundError are unrelated loading bugs." + "reason": "SAM3 vision encoder output attribute error and trust_remote_code cache FileNotFoundError are different problems." }, { "left": "issue:46143", "right": "issue:46144", - "accept": false, - "reason": "Both concern RoFormer, but kwargs passthrough and attention-interface usage are distinct concrete problems." + "accept": true, + "reason": "Both RoFormer issues point to the same underlying attention-path problem: methods bypassing the expected attention interface and not forwarding kwargs." }, { "left": "issue:44315", "right": "issue:46209", "accept": false, - "reason": "Liger Kernel model_init behavior and FP8 inference on B200 are unrelated." + "reason": "Liger-kernel model_init bug and FP8 inference failure on B200 are unrelated." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "FP8 save_pretrained on moe and FP8 inference support for qwen3-vl/qwen3.5 moe are different tasks." + "reason": "FP8 save_pretrained/moe behavior and FP8 inference support for Qwen3-VL/Qwen3.5-Moe are different issues." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Cannot import AutoImageProcessor and cannot import PreTrainedModel are separate export/import failures." + "reason": "AutoImageProcessor import failure and PreTrainedModel import failure are separate export problems." }, { "left": "issue:44485", "right": "issue:46664", "accept": false, - "reason": "GLM-5 RoPE implementation and video num_frames semantics are unrelated." + "reason": "GLM-5 RoPE implementation and SmolVLM num_frames semantics are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Generation cache missing during chunked_prefill and a typo in cache strategies are not the same bug." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A specific Qwen3.5-35B chat crash and a broad 'too many models break since 5.0' complaint are not the same underlying issue." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The summary stays conservative about a largely heterogeneous set, and the two accepted soft links are plausible as same-bug/mergeable pairs from the titles alone. No obvious overstatement beyond the evidence in the packet." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -18227,7 +18253,7 @@ "issue:45230", "issue:45278", "issue:45341", - "issue:45362", + "issue:45468", "issue:45478", "issue:45542", "issue:45753", @@ -18239,8 +18265,6 @@ "issue:46664" ], "soft_pairs": [ - "issue:43673|issue:44492", - "issue:45362|issue:45850", "issue:44273|issue:45542", "issue:44623|issue:45230", "issue:44360|issue:46664", @@ -18278,276 +18302,278 @@ "issue:31795|issue:33453", "issue:43576|issue:44273", "issue:44315|issue:44829", - "issue:43519|issue:46655" + "issue:43519|issue:46655", + "issue:44360|issue:45468", + "issue:44273|issue:44908" ], - "prepared_review_unit_hash": "5e4ec899703165fccb37dfc89d93b283cc8bd5c10849e2030f40c0066010a6e1", + "prepared_review_unit_hash": "45b944b6bf04f2bff0f0bcfae943d23cb82e55976c5d0e73f3eeeb884996b73b", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 40, - "serialized_chars": 39106, - "estimated_input_tokens": 9777, - "estimated_eval_tokens": 19810 + "serialized_chars": 39102, + "estimated_input_tokens": 9776, + "estimated_eval_tokens": 19808 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very noisy: most items are unrelated model-specific bugs, import/load regressions, or version-mismatch reports. I found no strong duplicate pairs worth merging.", - "confidence": 0.27, - "canonical_issue_reason": "issue:45850 is the broadest and most central report in the set, capturing the general \"v5 breaks many models\" regression theme that several other reports loosely orbit.", + "summary": "The cluster is noisy and mostly contains unrelated regressions. The strongest common thread is import/lazy-loading-related issues, but the provided pairs do not look like exact duplicates of the same concrete bug/change.", + "confidence": 0.46, + "canonical_issue_reason": "issue:44273 is the most representative broad bug in the set, centered on lazy loading/import behavior, which is the closest shared theme across the few related entries.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45850 is the best single representative issue for this noisy cluster, but it is still very vague and not a precise duplicate target.", + "best_issue_reason": "issue:44273 is the best global issue candidate because it names the underlying subsystem most clearly and could plausibly anchor nearby import/lazy-loading regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43673", - "right": "issue:44492", - "accept": false, - "reason": "Chunked-prefill cache regression vs a cache-strategy typo; related vocabulary only, not the same bug." - }, - { - "left": "issue:45362", - "right": "issue:45850", - "accept": false, - "reason": "Specific Qwen3.5 chat crash vs broad v5 breakage complaint; different symptoms and likely different code paths." - }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy-loading problem vs TensorBoard/TensorFlow backend import error; different subsystems." + "reason": "Both are import/runtime-related, but one is a lazy-loading problem while the other is a TF backend/tensorboard-specific error; not the same bug." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Specific processor.save_pretrained file-loss bug vs a generic bug report with no matching failure mode." + "reason": "One is a specific processor.save_pretrained missing-files bug; the other is an uninformative generic bug report." }, { "left": "issue:44360", "right": "issue:46664", "accept": false, - "reason": "DSA indexer discussion vs SmolVLM video num_frames semantics; unrelated topics." + "reason": "Completely different domains: DSA indexer/ReLU vs SmolVLM video padding/num_frames." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Both touch import behavior, but one is lazy loading and the other is unsafe sys.modules access; not the same underlying bug." + "reason": "Both touch import/lazy-loading behavior, but the concrete failures differ and there is not enough evidence of the same underlying defect." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Different Qwen3-family failures: one is a Qwen3-VL-Moe bug, the other is missing tensor-parallel plan support." + "reason": "Same model family, but one is a general Qwen3-VL-Moe bug and the other is a tensor-parallel _tp_plan issue." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Qwen3-VL bounding-box output issue vs PretrainedConfig type-checking regression; unrelated." + "reason": "Qwen3-VL output bug vs a PretrainedConfig type-checking regression; unrelated code paths." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Same model family, but one is a bbox-output problem and the other is an infinite-loop-on-image bug; different concrete failures." + "reason": "Both mention Qwen3-VL/Qwen3.5 vision models, but the symptoms and likely root causes are different." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-time slowdown from import * usage vs missing GenerationMixin cache in chunked_prefill; not the same bug." + "reason": "Import-time performance/lazy-loading is the only overlap; one is about import * usage, the other about chunked_prefill cache presence." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Broken transformers env command vs checkpoint reload regression after upgrade; different features." + "reason": "Both are v5 regressions, but in different surfaces: the env command vs checkpoint reload compatibility." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "Model-specific missing pad_token_id vs broad v5 breakage report; not a single concrete duplicate." + "reason": "A missing pad_token_id on one config is much more specific than the broad 'breaks too many models' complaint." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "Model-specific missing pad_token_idx vs broad v5 breakage report; too vague to be the same bug." + "reason": "A missing pad_token_idx on StableLmConfig is a specific config bug, not the same as a broad version-wide complaint." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both are upgrade/load problems, but one is general checkpoint reload failure and the other is remote-code/meta-init incompatibility; different causes." + "reason": "Both involve v5 upgrade/loading problems, but they affect different model families and likely different code paths." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "Cache-strategy typo vs broad v5 regression complaint; not the same underlying issue." + "reason": "A cache-strategy typo is unrelated to a generic version regression complaint." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "General import errors after upgrade vs broad \"too many models break\" report; overlap is too broad to treat as duplicate." + "reason": "Broad import errors after upgrade are not the same concrete issue as a general 'too many models' complaint." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 settings warning vs cache-strategy typo; unrelated." + "reason": "TF32 API settings and cache-strategy typo are unrelated." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Model-type mismatch warning vs Liger Kernel not applied with model_init; different bugs." + "reason": "Model-type warning vs Liger Kernel application bug are different problems." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import performance regression vs inability to import a specific Qwen3_5 module; shared import wording only." + "reason": "Both are import-related, but one is about import * and performance; the other is a missing module import for a specific class." }, { "left": "issue:44230", "right": "issue:46095", "accept": false, - "reason": "FP8 inference support request vs weight-conversion-script question; not the same change." + "reason": "FP8 inference support for Qwen3-VL/Moe is unrelated to a DeepSeek weight-conversion question." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Import takes long sometimes vs lazy loading not working properly; related area, but not enough evidence of one concrete bug." + "reason": "Both concern import/lazy-loading performance, but one is slow import timing and the other is a functional lazy-loading failure; not clearly the same bug." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Integration-test seeding problem vs scheduler kwargs being ignored; unrelated." + "reason": "Test seed flakiness and scheduler kwargs handling are unrelated." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "transformers env command issue vs scheduler kwargs regression; no shared code path." + "reason": "Different subsystems entirely: env command vs scheduler kwargs." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention 2 support error vs Qwen3-VL bounding-box hallucination; different model issues." + "reason": "Nemotron Flash Attention support and Qwen3-VL bounding-box output are unrelated." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version compatibility report vs Qwen3-VL bbox output bug; unrelated." + "reason": "Python-version compatibility and Qwen3-VL output issues are unrelated." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwargs bug vs PretrainedConfig type-checking regression; different components." + "reason": "Scheduler keyword handling is unrelated to PretrainedConfig type checking." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "position_ids buffer access error vs generic bug report; no duplicate evidence." + "reason": "Specific position_ids buffer access bug vs generic bug report." }, { "left": "issue:44492", "right": "issue:46328", "accept": false, - "reason": "Cache-strategy typo vs SequenceFeatureExtractor.pad numpy/list conversion issue; unrelated." + "reason": "Both are efficiency/cleanup issues, but one is a typo and the other is a padding conversion performance regression." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer class mismatch on save vs generic bug report; not a duplicate." + "reason": "Tokenizer_class mismatch in save_pretrained is specific; the other item is generic and non-specific." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor kwargs forwarding bug vs scheduler kwargs ignored; both mention kwargs, but on different APIs." + "reason": "Processor kwargs forwarding and scheduler kwargs forwarding are different APIs and bugs." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "General import slowdown vs cannot import HybridCache; similar symptom, different underlying failure." + "reason": "Both involve imports, but one is performance/lazy-loading and the other is a missing symbol export/import error." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-performance issue vs a small testing_utils bug; unrelated." + "reason": "Import-time performance issue and a testing_utils bug are unrelated." }, { "left": "issue:43976", "right": "issue:45850", "accept": false, - "reason": "Python version support regression vs broad v5 breakage complaint; not the same bug." + "reason": "Python version support regression is not the same as a broad 'many models break' complaint." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python version compatibility vs specific module import failure; unrelated." + "reason": "Python compatibility and a specific module import failure are unrelated." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Qwen3-VL bbox output bug vs Qwen3.5 MoE from_pretrained error; different model paths." + "reason": "Different Qwen3-family issues with different symptoms and model paths; not the same concrete bug." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "Optional post_init validation issue vs tied-weight-key AttributeError; different code paths." + "reason": "Optional post_init validation and tied-weights key handling are separate implementation issues." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion vs tokenizer loading regression; one is docs, the other is runtime behavior." + "reason": "Documentation confusion in model.forward is unrelated to tokenizer loading regression." }, { - "left": "issue:43576", + "left": "issue:41669", "right": "issue:44273", "accept": false, - "reason": "Broken env command vs lazy loading issue; different features and failure modes." + "reason": "Import * overhead and lazy-loading failure may be loosely related, but they are not clearly the same underlying change or bug." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel application bug vs flash_attention_3 degenerate training; unrelated training/runtime problems." + "reason": "Liger Kernel not applied with model_init and flash_attention_3 training degeneration are different issues." }, { "left": "issue:43519", "right": "issue:46655", "accept": false, - "reason": "Qwen3VL timestamp calculation vs SmolVLM video padding calculation; different processors and bugs." + "reason": "Different video-processor bugs in different models: timestamp calculation vs padding calculation." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer/ReLU discussion is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy-loading regression and inverse_sqrt scheduler kwargs are unrelated." } ] }, @@ -18582,6 +18608,7 @@ "issue:41669", "issue:42371", "issue:42913", + "issue:42936", "issue:43066", "issue:43298", "issue:43352", @@ -18596,9 +18623,7 @@ "issue:43717", "issue:44230", "issue:44263", - "issue:44273", "issue:44351", - "issue:44360", "issue:44485", "issue:44492", "issue:44623", @@ -18621,8 +18646,6 @@ "issue:46894" ], "soft_pairs": [ - "issue:44360|issue:45468", - "issue:44273|issue:44908", "issue:45468|issue:46894", "issue:45230|issue:45310", "issue:43576|issue:46218", @@ -18655,302 +18678,297 @@ "issue:31515|issue:38617", "issue:43572|issue:44861", "issue:43525|issue:44861", - "issue:28348|issue:43352", "issue:44230|issue:44623", + "issue:28348|issue:43352", "issue:29870|issue:31515", "issue:30333|issue:31515", "issue:31515|issue:36296", "issue:31515|issue:31795", "issue:44230|issue:45310", "issue:43298|issue:44230", - "issue:43717|issue:46144" + "issue:43717|issue:46144", + "issue:42936|issue:46018" ], - "prepared_review_unit_hash": "c1cce356c560ddbc654dc7cad481372648656938855384ae155a73a8614bd05d", + "prepared_review_unit_hash": "3265befcec5f5e61b44bd85dc3dd95a244ecc1c861db1d2924791abc1df50407", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 43, - "serialized_chars": 39627, - "estimated_input_tokens": 9907, - "estimated_eval_tokens": 20070 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 42, + "serialized_chars": 38806, + "estimated_input_tokens": 9702, + "estimated_eval_tokens": 19660 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly false-positive similarity matches across unrelated Transformer issues; no soft edge is strong enough to treat as a duplicate or same-change pair. The most concrete anchor is the Gemma-4 audio positional encoding report.", - "confidence": 0.87, - "canonical_issue_reason": "issue:45468 is the most concrete and central report in the set, with the strongest tie to other model-specific similarity matches and enough detail to serve as the cluster anchor.", + "summary": "The cluster is broadly heterogeneous, with a loose v5/regression theme but no true duplicate pairs among the soft edges. No PRs are present. If one issue must serve as the anchor, 43066 is the most concrete and representative of the tokenizer-v5 subset.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 43066 is the clearest concrete bug report here: a specific v5 tokenizer decoder-type regression.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45468 is the clearest actionable standalone bug report in the cluster: specific component, reproducible behavior, and active discussion.", + "best_issue_reason": "43066 is the most suitable representative because it is specific, actionable, and on the main recurring v5/tokenizer theme; most other issues are either vague placeholders or unrelated subsystems.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44360", - "right": "issue:45468", - "accept": false, - "reason": "Different bugs in different subsystems: DSA indexer/ReLU vs Gemma-4 audio positional encoding." - }, - { - "left": "issue:44273", - "right": "issue:44908", - "accept": false, - "reason": "One is a lazy-loading problem; the other is a scheduler kwarg propagation bug." - }, { "left": "issue:45468", "right": "issue:46894", "accept": false, - "reason": "Different models and failure modes: Gemma-4 audio encoding vs CohereASR training-loss shifting." + "reason": "Both are audio-related, but they describe different model bugs and different code paths." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "45230 is too generic to prove the same underlying bug as the specific Qwen3.5 MoE from_pretrained error." + "reason": "Both are generic loading failures, but one is a placeholder report and the other is a Qwen3.5 MoE-specific from_pretrained error." }, { "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "Different code paths: the env command vs an import failure for a specific model class." + "reason": "Both involve v5 breakage, but one is the `transformers env` command and the other is a missing model import symbol." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Unrelated model math issues: Qwen3VL timestamp calculation vs GLM-5 RoPE implementation." + "reason": "Processor timestamp calculation and GLM-5 RoPE implementation are different bugs in different components." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Both touch saving-related code, but one is pipeline save_pretrained and the other is tied-weight-key handling." + "reason": "Both touch save/pretrained internals, but Pipeline saving and tied-weight-key handling are distinct failures." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "No shared code-path; timestamp calculation and scheduler kwargs are separate concerns." + "reason": "Timestamp calculation in a processor and scheduler kwargs handling are unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-star cleanup performance issue vs a VRAM leak in Qwen3ForCausalLM are unrelated." + "reason": "Import-star cleanup/performance work is unrelated to a VRAM leak in Qwen3ForCausalLM." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion about forward args is not the same bug as model loading failing for Llama 3.2 Vision." + "reason": "Documentation confusion about inputs/past_key_values is not the same as a model-loading regression." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "Different model-specific bugs with no shared failing code path." + "reason": "These are different failures in different model implementations: memory leak versus RoPE math." }, { "left": "issue:46018", "right": "issue:46172", "accept": false, - "reason": "Both are deepseek-v4 related, but one is gating logic and the other is a tensor transpose issue." + "reason": "DeepSeek-V4 shared-expert gating and DeepSeek-V4 tensor transpose issues are distinct bugs." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Distinct implementation areas: RoPE math vs audio relative positional encoding." + "reason": "Different models and different implementation areas; not the same underlying defect." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 settings issue and VRAM leak are unrelated." + "reason": "TF32 API settings and a VRAM leak are unrelated issues." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "Offline-loading network leakage vs Qwen3VL timestamp math are separate bugs." + "reason": "Offline-loading network requests and processor timestamp math are unrelated." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "A broad v5 breakage complaint does not match a specific Qwen3_5 import error." + "reason": "A broad complaint that v5 breaks many models is not the same bug as a specific Qwen3_5 import failure." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Timestamp calculation and missing processor save files are different failure modes." + "reason": "Processor timestamp logic and missing save_pretrained files are different code paths." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return semantics in a GLM indexer is unrelated to Gemma-4 audio positional encoding." + "reason": "GLM MoE indexing and Gemma-4 audio positional encoding are unrelated bugs." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch is unrelated to Gemma-4 audio encoding." + "reason": "CPU-offload device mismatch and audio positional encoding are different problems." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "No common underlying code path between Qwen3VL processor math and VRAM leakage." + "reason": "Processor timestamp calculation and VRAM leak symptoms are unrelated." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Audio continuous batching support and DeepSeek-V4 weight conversion are different change requests." + "reason": "Continuous batching/audio support and DeepSeek-V4 weight conversion are not the same change." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash Attention support for Nemotron and a Qwen3_5 import failure are unrelated." + "reason": "Flash Attention support for NemotronH and a Qwen3_5 import error are different issues." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "Missing HybridCache export and missing Qwen3_5ForCausalLM export are different symbols and likely different causes." + "reason": "Both are import errors, but they involve different missing symbols and different modules." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both concern tokenizer v5 behavior, but one is a broad regression report and the other a specific decoder-type mismatch; not enough to merge." + "reason": "Both are tokenizer v5 regressions, but one is a general behavior change and the other is a specific decoder-type mismatch; not clearly one bug." }, { "left": "issue:28348", "right": "issue:45250", "accept": false, - "reason": "Both mention Flash Attention 2.0, but one is GPT2LMHeadModel support and the other is a vague generic report." + "reason": "They only share the Flash Attention 2.0 wording; one is GPT2 support and the other is a generic title." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs propagation and a nonsense/unemployment issue are unrelated." + "reason": "Scheduler kwargs handling is unrelated to the placeholder issue title." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Typo in cache strategies does not match the unrelated placeholder issue." + "reason": "A typo in cache strategies is unrelated to the placeholder issue." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the placeholder issue." + "reason": "CLIPOutput attention assignment regression is unrelated to the placeholder issue." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "Placeholder issue cannot be matched to a broad complaint about v5 breakages." + "reason": "A placeholder title does not indicate the same underlying bug as the broad v5-breakage complaint." }, { "left": "issue:44938", "right": "issue:44998", "accept": false, - "reason": "Python 3.14 load failure is unrelated to the placeholder issue." + "reason": "Python 3.14 loading failure is unrelated to the placeholder issue." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weight key handling and chat-template tool-call crashing are different code paths." + "reason": "Tied-weight key crashes and chat-template tool-call crashes are different code paths." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow from_pretrained loading is not the same as the missing layer_type_validation import error." + "reason": "Slow from_pretrained loading and a missing import symbol are unrelated." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_idx in a config is unrelated to tied-weight-key attribute handling." + "reason": "StableLmConfig pad-token regression and tied-weight-key handling are different bugs." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is a config bug, not tied-weight-key crashing." + "reason": "Llama4Config pad_token_id absence is unrelated to tied-weight-key crashes." }, { - "left": "issue:28348", - "right": "issue:43352", + "left": "issue:44230", + "right": "issue:44623", "accept": false, - "reason": "Both mention Flash Attention 2.0, but they target different models and different support gaps." + "reason": "FP8 inference support and missing processor save files are unrelated." }, { - "left": "issue:44230", - "right": "issue:44623", + "left": "issue:28348", + "right": "issue:43352", "accept": false, - "reason": "FP8 Qwen3-VL/MoE support request and processor save-pretrained missing files are unrelated." + "reason": "Both mention Flash Attention 2.0, but they concern different models and different symptoms." }, { "left": "issue:29870", "right": "issue:31515", "accept": false, - "reason": "BART generate() decoder mask bug and slow checkpoint loading are separate issues." + "reason": "BART generate mask error and slow checkpoint loading are unrelated." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLFlow reporting on training failure is unrelated to checkpoint loading speed." + "reason": "MLFlow reporting failure and slow loading are unrelated issues." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Slow loading and tensor-parallel training bug are different code paths." + "reason": "Checkpoint loading performance and tensor-parallel training are different problems." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Model loading performance and documentation confusion are not the same bug." + "reason": "Loading performance and documentation confusion do not describe the same bug." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 inference support for Qwen3-VL/Qwen3.5 and a from_pretrained import error are different issues." + "reason": "FP8 support request and Qwen3.5 MoE loading error are different concerns." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Audio continuous batching and FP8 support are unrelated feature gaps." + "reason": "Continuous batching for audio models and FP8 inference support are unrelated." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba init_weights behavior and RoFormer attention-interface usage are unrelated implementation issues." + "reason": "Mamba init-weights behavior and RoFormer attention interface usage are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2former ignore_value handling and DeepSeek-V4 shared-expert gating are distinct bugs." } ] }, @@ -18980,11 +18998,9 @@ "issue:43668", "issue:43813", "issue:44031", - "issue:46018", "issue:46144" ], "soft_pairs": [ - "issue:42936|issue:46018", "issue:43668|issue:44031", "issue:42936|issue:43240", "issue:43240|issue:46144", @@ -18996,100 +19012,94 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "7b2ee3e371b1e682705c0db04cb2ac808ab613a27f44a96d44451f1400a900a6", + "prepared_review_unit_hash": "01756314d74ff7fd2bafa898ce619016749025cbeb37382eaabab76eeb6ef34c", "component_budget": { "node_count": 502, "item_count": 502, - "soft_pair_count": 1846, - "serialized_chars": 957673, - "estimated_input_tokens": 239419, - "estimated_eval_tokens": 479094 + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 }, "budget": { - "node_count": 8, - "item_count": 8, - "soft_pair_count": 11, - "serialized_chars": 7774, - "estimated_input_tokens": 1944, - "estimated_eval_tokens": 4144 + "node_count": 7, + "item_count": 7, + "soft_pair_count": 10, + "serialized_chars": 6932, + "estimated_input_tokens": 1733, + "estimated_eval_tokens": 3722 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are not duplicates of one another; they span unrelated models and bugs (normalization, loss kwargs, tokenizer regex warnings, expert gating, attention interface, etc.). All soft-similarity links should be rejected.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue: this set is heterogeneous and does not form a single duplicate bug cluster.", + "summary": "This cluster is not a duplicate set: the issues span unrelated models/components (norm behavior, Mask2Former loss handling, ModernBERT typing, tokenizer warnings, RoFormer attention plumbing, etc.). None of the soft pairs look like the same underlying bug.", + "confidence": 0.98, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good global representative because the items describe different underlying problems in different code paths.", + "best_issue_reason": "No single canonical issue fits this cluster because the items are unrelated; if one must be highlighted, issue:43240 is the most concrete and broadly discussed bug report.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42936", - "right": "issue:46018", - "accept": false, - "reason": "Mask2Former ignore_value handling and DeepSeek-V4 shared-expert gating are unrelated model bugs." - }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "One is a type-hint mistake in ModernBERT config; the other is a tokenizer regex warning regression." + "reason": "Different problems: a type-hint bug in ModernBERTConfig versus a tokenizer regex warning regression." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2Former ignore_value usage and missing kwargs propagation to loss calculation are different failure modes." + "reason": "Mask2Former ignore_value handling is unrelated to kwargs not being forwarded into a loss function." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss kwargs forwarding is unrelated to RoFormer's attention-interface implementation issue." + "reason": "Loss kwarg forwarding and RoFormer attention interface usage are separate code paths and bug types." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "These affect different subsystems: Mask2Former loss masking vs tokenizer warning behavior." + "reason": "Mask2Former loss/ignore_value behavior does not match tokenizer regex warning behavior." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "Mask2Former ignore_value and RoFormer attention interface do not describe the same bug." + "reason": "Mask2Former ignore_value logic and RoFormer attention interface wiring are unrelated model bugs." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "A missing use of ignore_value is unrelated to a typo in a quantization operation name." + "reason": "A missing ignore_value use is unrelated to a typo in a quantization attribute name." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Passing kwargs to a loss function and a wrong type hint in config are separate issues." + "reason": "Passing kwargs to loss calculation is unrelated to an incorrect type hint for norm_eps." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "ModernBERT norm_eps typing and a quantization-operation typo are unrelated code changes." + "reason": "ModernBERTConfig typing and a quantization-attribute typo are different issues with no shared code-path." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Llama4TextL2Norm RMS-norm question and ModernBERT norm_eps type hint are different model/config concerns." + "reason": "Llama4TextL2Norm vs RMS norm semantics is a model implementation question, not a ModernBERTConfig type-hint bug." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Mask2Former ignore_value usage and ModernBERT config typing are not the same underlying bug." + "reason": "Mask2Former ignore_value handling and ModernBERTConfig norm_eps typing are unrelated." } ] }, @@ -19122,7 +19132,7 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "d8f65d2c36de5313e6a37b644443113be12f49c5d960ecad943626d5e783942c", + "prepared_review_unit_hash": "1b7be70740fefbce69d0859117844a270100235e1da5d3cd02219b9791650827", "component_budget": { "node_count": 3, "item_count": 3, @@ -19144,30 +19154,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The three issues are about unrelated topics: tokenizer truncation behavior, ReduceLROnPlateau-like scheduling, and Wav2Vec2 word timestamps. This is not a duplicate cluster.", - "confidence": 0.99, - "canonical_issue_reason": null, + "summary": "These three issues are not duplicates: they concern unrelated topics (tokenizer truncation behavior, a ReduceLROnPlateau-like scheduler feature, and Wav2Vec2 word boundary offsets).", + "confidence": 0.98, + "canonical_issue_reason": "Issue #10519 is the most concrete, self-contained feature request in the set; the other two are unrelated problems and do not form a single underlying bug/feature.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good canonical representative because the items are unrelated feature requests/questions rather than the same underlying bug.", + "best_issue_reason": "#10519 is the best representative only in the sense that it is the clearest and most specific issue, but the cluster does not appear to be a true duplicate group.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Different domains and problems: truncation strategy in tokenization vs extracting word-level time offsets in Wav2Vec2." + "reason": "Different areas and requests: tokenizer truncation direction vs. Wav2Vec2 word timestamp offsets. No shared underlying bug or change." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "Unrelated requests: truncation-from-beginning option vs ReduceLROnPlateau-like learning-rate scheduling." + "reason": "Token truncation behavior is unrelated to ReduceLROnPlateau-like scheduling functionality." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "Completely different functionality: LR scheduler behavior vs Wav2Vec2 timestamp offsets; not the same bug or change." + "reason": "A learning-rate scheduler feature request and Wav2Vec2 word offset extraction are unrelated." } ] }, @@ -19199,12 +19209,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "d24a3da6d27b5d50861d38c77f2318d4117b66e75fb33b03c6b0e8392d695e49", + "prepared_review_unit_hash": "38bf940c4f209755e2acdf93dabff257fb52cd5ea7a99afcc7be398aaeab97fe", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -19212,7 +19222,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -19221,24 +19231,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues are not duplicates: one is about Tapas tokenization mismatch, one about an authentication/error-message problem for private model downloads, and one about failure to load bert-base-uncased. The only overlap is a broad similarity around model/tokenizer loading, but they describe different failure modes and likely different fixes.", - "confidence": 0.9, - "canonical_issue_reason": "Issue #16618 is the broadest and most user-facing of the three, centered on an actual model-load failure; the others are narrower or about a different subsystem (tokenization vs. auth/error messaging).", + "summary": "These are not true duplicates: one is about TAPAS tokenization, one about private-model download/auth error messaging, and one about a generic BERT loading failure. The similarities are superficial and point to different code paths.", + "confidence": 0.87, + "canonical_issue_reason": "Issue #16618 is the broadest and most central user-facing failure in the set, so it works best as the anchor even though it is not actually a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the best representative issue because it captures the generic model-loading symptom more directly than the others and is the most likely to be reused as a reference point for related load failures.", + "best_issue_reason": "#16618 is the most generally phrased and likely most reusable report among the three, but it still describes a different problem than the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tapas tokenization mismatch is a tokenizer/parity issue, while 'Can't load the model for bert-base-uncased' is a model-loading failure; they do not look like the same underlying bug or the same fix." + "reason": "TAPAS tokenization mismatch is a model-specific preprocessing issue, not the same bug as a generic \"can't load bert-base-uncased\" failure." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "#15980 is specifically about a bad error message when downloading a private model without login, which is an authentication/error-reporting issue, not the same underlying problem as failing to load bert-base-uncased." + "reason": "Private-model download/auth error messaging is about access-control handling, while #16618 is a general model-loading problem for a public model." } ] }, @@ -19268,7 +19278,7 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "2fe60beceb37ed22a00c474d88661a18e6958a51e66adfd0871a933f0c6f7c4f", + "prepared_review_unit_hash": "90e87c729a81dd1f09e3b5ddc072c4bb0c5f03d5f26cee537a643ffe876d8212", "component_budget": { "node_count": 2, "item_count": 2, @@ -19290,18 +19300,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates: one asks for Flash Attention 2 support on more architectures, the other for GGUF support on more architectures. They are separate feature requests in the same broad area, so the soft match should be rejected.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue here; the issues target different features and different code paths (Flash Attention 2 vs GGUF support).", + "summary": "These are two separate feature requests in the same general area (community support for additional model/attention architectures), but they target different capabilities and different code paths: Flash Attention 2 vs GGUF. They should not be deduplicated.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:26350 is the better representative thread only because it has higher discussion activity and more traction, but it is not a duplicate of #33260.", + "best_issue_reason": "No single canonical issue fits both, since the requests are about different features; if one must be chosen as a representative issue, issue 26350 is slightly more central as an earlier, specific architecture-support request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different requested capabilities: Flash Attention 2 architecture support versus GGUF architecture support. Same broad subsystem, but not the same underlying bug or change." + "reason": "Different underlying requests: one is for Flash Attention 2 support, the other for GGUF support. They touch different integration paths and are not the same bug or change." } ] }, @@ -19366,12 +19376,12 @@ "issue:46222|issue:46931", "issue:46226|issue:46931" ], - "prepared_review_unit_hash": "57a5c0ede16021af9ea9beab905ac8b2bf848d8f9845cd5c4198ab3b267c9f24", + "prepared_review_unit_hash": "88e0a639e03a60463eb0cea23db487aeb708e89788e6a7f92aa8185b4849e3bc", "component_budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 29, - "serialized_chars": 15984, + "serialized_chars": 15981, "estimated_input_tokens": 3996, "estimated_eval_tokens": 8248 }, @@ -19379,7 +19389,7 @@ "node_count": 9, "item_count": 9, "soft_pair_count": 29, - "serialized_chars": 15984, + "serialized_chars": 15981, "estimated_input_tokens": 3996, "estimated_eval_tokens": 8248 }, @@ -19388,195 +19398,195 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Most items are not true duplicates: the Turkish, Spanish, and Romanian translation requests are separate language-specific issues. The only clear duplicate subgroups are the three identical placeholder i18n template issues, and the two identical Thai-titled issues.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 43834 is the oldest of the identical placeholder translation-template issues and is the best representative of that duplicate subcluster.", + "summary": "This cluster is mostly a mix of unrelated language-specific translation requests plus one generic placeholder-template subgroup and one separate Thai duplicate pair. The safest duplicate calls are the template-title issues ([i18n-] Translating docs to ) and the identical Thai-title pair; the Turkish, Spanish, and Romanian requests should stay separate because they target different languages.", + "confidence": 0.88, + "canonical_issue_reason": "43834 is the cleanest representative of the generic placeholder-template subgroup: it is the earliest of those issues and serves as the best anchor for that duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "43834 is the clearest representative of the template-duplicate subgroup; the other concrete language requests are distinct tasks, not duplicates.", + "best_issue_reason": "43834 is the best overall issue representative for the main duplicate subgroup; the other high-similarity items are either the same template issue or separate language requests that should not be merged.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different language-specific translation requests; same i18n theme is too broad to make them duplicates." + "reason": "Different language-specific translation request versus the generic placeholder template; the similarity is thematic, not the same issue." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Turkish translation request vs generic placeholder translation issue; not the same underlying task." + "reason": "Turkish request and generic template issue are not the same underlying task." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Concrete Turkish docs translation request is distinct from the placeholder template issue." + "reason": "Turkish-specific translation request versus template placeholder; not a duplicate." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Different target languages (Turkish vs Romanian), so separate issues." + "reason": "Different target languages, so these are separate translation requests." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different target languages (Turkish vs Spanish), so not duplicates." + "reason": "Turkish and Spanish translation requests are distinct." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish translation request is not the same as the generic placeholder issue." + "reason": "Spanish-specific request versus the generic template issue; not the same underlying issue." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Spanish translation request vs placeholder template issue; too different to merge." + "reason": "Different language request and template placeholder; similarity is too broad to merge." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Different target languages (Spanish vs Romanian), so separate tasks." + "reason": "Spanish and Romanian translation requests are separate." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Spanish translation request vs generic placeholder issue; not a duplicate." + "reason": "Different language request versus template placeholder; not a duplicate." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian translation request is distinct from the generic placeholder issue." + "reason": "Romanian request versus generic placeholder template; only the issue pattern overlaps." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Romanian translation request vs placeholder template issue; not the same underlying issue." + "reason": "Different concrete language request; the shared structure is not enough to merge." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian translation request is separate from the placeholder translation-template issue." + "reason": "Romanian-specific request versus template placeholder; not the same issue." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Different content entirely; this is a Romanian docs request, not the Thai-titled issue." + "reason": "Different language/request content; the similarity is only at a broad template level." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Different content entirely; this is a Romanian docs request, not the Thai-titled issue." + "reason": "Different language/request content; not a duplicate." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "Same generic placeholder i18n translation issue; titles/template are effectively identical." + "reason": "Same generic placeholder-style title and very high similarity signals; body text is not shown, so this is a likely duplicate rather than a certain exact match." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Same generic placeholder i18n translation issue; no distinguishing details indicate separate tasks." + "reason": "Same generic placeholder-style title and very high similarity signals; likely duplicate based on the available metadata." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "The placeholder translation-template issue and the Thai-titled issue are not the same underlying task." + "reason": "The titles and apparent language context do not line up; the shared similarity is not enough to treat these as the same issue." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "The placeholder translation-template issue and the Thai-titled issue are not the same underlying task." + "reason": "Only template-level similarity is visible; not enough evidence that these are the same request." }, { "left": "issue:43834", "right": "issue:46931", "accept": false, - "reason": "Different titles and likely different content; not the same duplicate issue." + "reason": "Different title and likely different underlying request; not a duplicate." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "Same placeholder translation-template issue; effectively identical report." + "reason": "Same placeholder-style title and extremely high similarity signals; likely duplicate, though the body text is not exposed here." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Placeholder translation template vs Thai-titled issue; too different to be duplicates." + "reason": "Different language/request content; template similarity alone is not enough." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Placeholder translation template vs Thai-titled issue; too different to be duplicates." + "reason": "Different language/request content; not clearly the same issue." }, { "left": "issue:44365", "right": "issue:46931", "accept": false, - "reason": "Different Thai-language titles/content; not the same underlying issue." + "reason": "Different title/content and no clear duplicate relationship." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Generic placeholder translation issue vs Thai-titled issue; not a duplicate." + "reason": "Template similarity only; the visible metadata does not support the same underlying request." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Generic placeholder translation issue vs Thai-titled issue; not a duplicate." + "reason": "Template similarity only; not enough evidence to merge." }, { "left": "issue:44370", "right": "issue:46931", "accept": false, - "reason": "Different Thai-language titles/content; not the same underlying issue." + "reason": "Different issue content; not a duplicate." }, { "left": "issue:46222", "right": "issue:46226", "accept": true, - "reason": "Identical Thai-titled issue reports with matching content; likely exact duplicates." + "reason": "Same exact Thai title and very high similarity signals; likely duplicate based on metadata, though the body text is not shown." }, { "left": "issue:46222", "right": "issue:46931", "accept": false, - "reason": "Different Thai titles/content; no evidence they describe the same task." + "reason": "Different titles and no strong evidence they describe the same request." }, { "left": "issue:46226", "right": "issue:46931", "accept": false, - "reason": "Different Thai titles/content; no evidence they describe the same task." + "reason": "Different titles and no strong evidence of a duplicate relationship." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The verdicts are conservative and grounded in the packet. The three placeholder i18n template issues share the same title and are plausibly the same report cluster, with 43834 correctly chosen as the oldest representative. The Thai-titled pair (46222/46226) has identical titles and matching body lengths, so treating them as duplicates is supported. The language-specific Turkish, Spanish, and Romanian issues are reasonably kept separate based on their distinct titles. " + "accept": false, + "feedback": "The clustering is broadly plausible, but the writeup is a bit too strong for the packet. The packet shows identical or very similar titles, body lengths, and soft-similarity signals, but it does not expose the body text, so phrases like \"exact duplicate\" and \"near-identical body\" go beyond what is directly supported. The conservative parts are fine, especially separating the clearly different language requests, but the accepted duplicate verdicts should be phrased as likely duplicates rather than certain exact matches." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -19656,274 +19666,274 @@ "issue:42503|issue:46653", "issue:45027|issue:46653" ], - "prepared_review_unit_hash": "05e6e39bd301f2dacf782e016f09d8fe5aff1b805112619e2613f920ec250693", + "prepared_review_unit_hash": "16e60a44c56da7a4bbce2e198033acf1239d33e26d05d0d5b370a59dedf8da9e", "component_budget": { "node_count": 19, "item_count": 19, "soft_pair_count": 40, - "serialized_chars": 24673, - "estimated_input_tokens": 6169, - "estimated_eval_tokens": 12594 + "serialized_chars": 24672, + "estimated_input_tokens": 6168, + "estimated_eval_tokens": 12592 }, "budget": { "node_count": 19, "item_count": 19, "soft_pair_count": 40, - "serialized_chars": 24673, - "estimated_input_tokens": 6169, - "estimated_eval_tokens": 12594 + "serialized_chars": 24672, + "estimated_input_tokens": 6168, + "estimated_eval_tokens": 12592 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a loose bag of distinct model-request issues (vision, BERT variants, audio/TTS/ASR, OCR, etc.). The soft edges are title/body similarity only; none look like the same underlying bug or change.", - "confidence": 0.88, - "canonical_issue_reason": "Pick issue 36879 ('Add RF-DETR model') as the cleanest generic model-add request exemplar, but it does not duplicate the rest of the cluster.", + "summary": "These items are mostly unrelated model-request issues. Despite title-level similarity, they do not share the same concrete bug or code change, so no soft edge should be accepted.", + "confidence": 0.98, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 36879 is the most canonical representative of the cluster\u2019s general 'add a model' pattern; the other issues are different models or modalities, so none are better duplicate representatives.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36718", "right": "issue:41115", "accept": false, - "reason": "RoMa keypoint matcher vs MiniCPM3 architecture request: different model families and different implementation surfaces." + "reason": "Different domains: keypoint matching vs a MiniCPM3 architecture request." }, { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBert config vs DEIMv2: unrelated model additions, not the same change." + "reason": "Unrelated changes: EuroBert config entry vs a DEIMv2 model addition." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config vs Qwen3-TTS support: different models and modalities." + "reason": "Different capabilities: EuroBert config vs Qwen3-TTS support." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert config vs RF-DETR model: both are model requests, but for different concrete models and code paths." + "reason": "Different model families: EuroBert config vs RF-DETR model add." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR vs DEIMv2 are distinct vision model requests, not a duplicate report." + "reason": "Both are model additions, but for different vision models and code paths." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR model vs LFM2.5 Audio 1.5B: different domains and implementations." + "reason": "Detection model request vs audio model request; not the same change." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "RF-DETR vs weight-sparse transformer: unrelated architectures and change scopes." + "reason": "Different model families: RF-DETR vs weight-sparse transformer support." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR model vs Qwen3-ASR support: different models and different modalities." + "reason": "Vision detection vs speech ASR; unrelated underlying change." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR vs Qwen3-TTS support: unrelated model additions." + "reason": "Different modalities and models: RF-DETR vs Qwen3-TTS." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT vs Qwen3-ASR: different model families and different code paths." + "reason": "NeoBERT and Qwen3-ASR are unrelated model additions." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT vs weight-sparse transformer: not the same model or feature." + "reason": "BERT-family model request vs weight-sparse transformer request." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT vs LFM2.5 Audio 1.5B: unrelated text vs audio model requests." + "reason": "Text LM request vs audio model request." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR model vs NeoBERT: separate model-addition requests." + "reason": "Different model additions; no shared code path or exact target." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT vs ModernVBERT models: both are model additions, but for different architectures." + "reason": "NeoBERT vs ModernVBERT are separate model integrations." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture vs Qwen3-ASR support: different models and modalities." + "reason": "Architecture request vs TTS support; not the same bug/change." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 vs weight-sparse transformer: not the same underlying change." + "reason": "MiniCPM3 architecture vs sparse-transformer support are different changes." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 architecture vs ModernVBERT models: different model families." + "reason": "Different model families and integration points." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 architecture vs Voxtral TTS support: unrelated request scopes." + "reason": "MiniCPM3 architecture request vs Voxtral TTS support." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 vs Qwen3-TTS support: distinct model requests, not duplicates." + "reason": "Vision detection request vs ASR support; different pipeline." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models vs Qwen3-ASR support: different model families and tasks." + "reason": "ModernVBERT vs ASR are unrelated model requests." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT vs Qwen3-TTS: unrelated additions." + "reason": "ModernVBERT config/support vs Qwen3-TTS; different models and modalities." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT models vs BERT-like models with RoPE: related naming, but not the same concrete model or change." + "reason": "Generic BERT-RoPE support vs ModernVBERT model addition." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like models with RoPE vs Voxtral TTS support: different model classes and modalities." + "reason": "BERT/RoPE support vs Voxtral TTS; unrelated." }, { "left": "issue:42738", "right": "issue:46653", "accept": false, - "reason": "BERT-like models with RoPE vs 'Mirko Privitera' is unrelated; the latter is not the same kind of model request." + "reason": "Model request vs unrelated person-named issue; no common change." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS support vs Qwen3-ASR support: same product line, but different concrete capabilities and implementation." + "reason": "Qwen3-TTS and Qwen3-ASR are different modalities and likely separate implementations." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS support vs weight-sparse transformer: unrelated code paths." + "reason": "Two different model requests with different integration points, not one concrete change." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS support vs Voxtral-4B-TTS-2603: both audio-related, but different models and integration work." + "reason": "TTS support vs Voxtral TTS model addition; different models and code paths." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS support vs DeepSeek-OCR2: different modalities and model families." + "reason": "TTS support vs OCR model addition; unrelated." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR support vs LFM2.5 Audio 1.5B: audio-related, but separate model requests with different changes." + "reason": "ASR support vs audio model request; different model integrations." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR support vs EUPE: unrelated model types and implementation surfaces." + "reason": "Qwen3-ASR support vs EUPE; unrelated model families." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR support vs weight-sparse transformer: different tasks and code paths." + "reason": "ASR support vs sparse-transformer request; different changes." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio 1.5B vs SAM3-LiteText: different modality and model request." + "reason": "Audio model vs SAM3-LiteText; not the same bug/change." }, { "left": "issue:43909", "right": "issue:46645", "accept": false, - "reason": "LFM2.5 Audio 1.5B vs SWA+sink option for Granite models: one is a new audio model, the other is a model option change." + "reason": "Audio model vs SWA+sink Granite option; different model families." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer request vs EUPE: different model families and different changes." + "reason": "Sparse transformer feature vs EUPE model request." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR support vs Voxtral TTS support: both speech-related, but not the same concrete fix." + "reason": "ASR model request vs Voxtral TTS; different modalities and implementations." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT vs Voxtral TTS support: unrelated model additions." + "reason": "NeoBERT vs Voxtral TTS; unrelated model additions." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR support vs DeepSeek-OCR2: different modalities and code paths." + "reason": "ASR support vs DeepSeek-OCR2; separate integrations." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE vs TIPSv2: both vision-language related, but distinct model requests and implementation work." + "reason": "EUPE and TIPSv2 are distinct model requests." }, { "left": "issue:42503", "right": "issue:46653", "accept": false, - "reason": "ModernVBERT models vs 'Mirko Privitera' are unrelated; this is not the same underlying issue." + "reason": "Model request vs unrelated person issue." }, { "left": "issue:45027", "right": "issue:46653", "accept": false, - "reason": "Voxtral TTS support vs 'Mirko Privitera' are unrelated; no duplicate relationship." + "reason": "TTS model request vs unrelated person issue." } ] }, @@ -19953,7 +19963,7 @@ "soft_pairs": [ "issue:38740|issue:46443" ], - "prepared_review_unit_hash": "f333b8ca1a15ab91f3cefd8308d79051954f78682ed0de620b154aee06cca1c8", + "prepared_review_unit_hash": "5cd8c890d4bcc179e9c819b1ea762486136795be4df327e465181721a0451803", "component_budget": { "node_count": 2, "item_count": 2, @@ -19975,18 +19985,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster looks spurious: one item is a docs request to add Pruna, while the other is an unrelated Vision RoPE refactor. They do not describe the same underlying change or bug.", + "summary": "The two issues are unrelated: one is a documentation request for adding Pruna as an optimization framework, while the other is a Vision RoPE refactor. They should not be clustered as duplicates or the same underlying task.", "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue #46443 is the more substantive technical change request; #38740 is a separate documentation enhancement and not a duplicate candidate.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38740", "right": "issue:46443", "accept": false, - "reason": "Different topics and code paths: documentation about adding Pruna versus a Vision RoPE refactor. No evidence they refer to the same bug or change." + "reason": "Different topics and likely different workstreams: docs/framework listing vs. vision model positional-encoding refactor. No evidence they refer to the same bug or change." } ] }, @@ -20016,12 +20026,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "f255f8c9c9a1ef2690deab972265aff5696ec2d9dcfdc3b289b0cee396555de5", + "prepared_review_unit_hash": "2e017f9c21fcf03e45dba68ae8edfb487c99f75213210b9e0ad96076db556603", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, + "serialized_chars": 1479, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -20029,7 +20039,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, + "serialized_chars": 1479, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -20038,18 +20048,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates: one is about set block decoding, the other is a request for DeepSeek v3.2 support. They describe different product concerns with no shared code-path evidence.", + "summary": "The two issues are about different topics: one is about set block decoding, the other about DeepSeek v3.2 support. They do not appear to describe the same underlying problem or change.", "confidence": 0.97, - "canonical_issue_reason": "Issue 43037 is the more concrete user-facing request (model support), but it is not the same underlying problem as 41084.", + "canonical_issue_reason": "No true canonical issue: these are unrelated requests. If forced to pick a representative, issue 43037 is the clearer standalone feature request, but it is not a duplicate of 41084.", "canonical_pr_reason": null, - "best_issue_reason": "43037 is the better standalone issue because it is specific and actionable, whereas 41084 is a separate decoding-related topic.", + "best_issue_reason": "Issue 43037 is the better standalone representative because it is a clearly scoped support request, but it should not be treated as the same issue as 41084.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different subjects: decoding behavior versus adding support for a specific model version. No evidence they describe the same bug or change." + "reason": "Different subjects and likely different code paths; no evidence they are the same bug or change." } ] }, @@ -20079,40 +20089,40 @@ "soft_pairs": [ "issue:41306|issue:41826" ], - "prepared_review_unit_hash": "51613b6e0dd5e01bad415b5aa4c4a8cf5502db4344edee3ba448717522ef573b", + "prepared_review_unit_hash": "5462f526d02d5e0e3ddbd4ee8e2a8198994a2e4eea164796fb0c2c215df0d4dc", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1585, - "estimated_input_tokens": 397, - "estimated_eval_tokens": 1050 + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1585, - "estimated_input_tokens": 397, - "estimated_eval_tokens": 1050 + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one asks for a `from_pretrained()` flag for TiledFusedLogitsLoss / FLCE, while the other is a feature request to integrate TiledMLP for lower memory use. They share only the general 'Tiled' naming.", + "summary": "These are both Tiled-related feature requests, but they target different components and different changes: one asks for a `from_pretrained()` flag for TiledFusedLogitsLoss/LigerFusedLinearCrossEntropy, the other proposes integrating TiledMLP for lower memory usage. They do not look like duplicates.", "confidence": 0.97, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue 41306 is the more concrete, narrowly scoped API request; it is easier to treat as the representative item for this small cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 41306 is the more concrete, narrowly scoped request, but it is not the same underlying problem as 41826.", + "best_issue_reason": "41306 is the best single issue to keep as the cluster representative because it is a specific request with a clear entry point (`from_pretrained()`), while 41826 is a separate feature proposal.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41306", "right": "issue:41826", "accept": false, - "reason": "Different components and goals: FLCE flag handling in `from_pretrained()` versus adding TiledMLP for memory reduction. Shared wording is too superficial to treat as the same issue." + "reason": "Reject: they concern different features and code paths. 41306 is about a FLCE/TiledFusedLogitsLoss `from_pretrained()` flag, while 41826 is about adding TiledMLP for memory reduction; same subsystem name is not enough to make them the same underlying issue." } ] }, @@ -20142,12 +20152,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "ee31896ff54927a65d8f2406812394f94d3c5d2d5ef7acfec3b327720b85a3a7", + "prepared_review_unit_hash": "5b1743bdf6fa6ab995b85eda40351d80651905dc883fbb2107b689a8d35203cb", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1583, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -20155,7 +20165,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1583, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -20164,18 +20174,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two distinct Trainer-related issues: one about making `remove_unused_column` more customizable, and one about FLOPS estimation for custom trainers. They do not appear to describe the same underlying bug or change.", - "confidence": 0.92, - "canonical_issue_reason": "Issue #43598 is the broader, more central Trainer configurability topic and is more likely to serve as the representative issue for this small cluster.", + "summary": "Both issues are about custom trainers in `Trainer`, but they target different behaviors: one is about `remove_unused_column` customizability, the other about FLOPS estimation. They are related in theme, not the same underlying bug.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43598 is the broader, more representative Trainer customizability problem and best captures the cluster\u2019s theme.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #43598 has the clearer, more general problem statement; #43600 is a narrower metric-estimation concern.", + "best_issue_reason": "Issue 43598 is the best overall representative because it is the more general Trainer customization concern, while 43600 is a narrower metrics-specific case.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Different problems: one is about `remove_unused_column` behavior/customization, the other about estimating FLOPS for custom trainers. They share the Trainer subsystem but not the same code-path change." + "reason": "They share the broad custom-trainer context, but the concrete problems differ: one concerns `remove_unused_column` behavior, the other FLOPS counting. These would not naturally be fixed by one patch." } ] }, @@ -20205,7 +20215,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "80c4bdb724446888f07f1575f267e7723cf865f1791a83fc98d3de194f88ab00", + "prepared_review_unit_hash": "d1d71c24d3ea322e6b91dd260f6631d1650e883cb35da46a39ceefe32ef3caca", "component_budget": { "node_count": 2, "item_count": 2, @@ -20227,18 +20237,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one requests AutoModel/AutoModelForConditionalGeneration registration for Qwen3-Omni, while the other requests AutoModelForSequenceClassification support for Qwen3.5. They involve different model families and different heads/tasks.", - "confidence": 0.94, - "canonical_issue_reason": "No canonical duplicate issue here; 43684 targets Qwen3-Omni registration for generic generation classes, while 44405 targets Qwen3.5 sequence-classification support.", + "summary": "These are not duplicates: one requests Qwen3-Omni auto-model registration, while the other requests AutoModelForSequenceClassification support for Qwen3.5. They concern different model families and different auto-class paths.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "If a single representative is needed, 43684 is the broader model-registration request, but it should not be treated as the same bug as 44405.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different underlying change requests: Qwen3-Omni AutoModel registration vs Qwen3.5 AutoModelForSequenceClassification support. Different model/config and different task head, so not the same bug." + "reason": "Different underlying gaps: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 sequence-classification support. Not the same code path or mergeable as one fix." } ] }, @@ -20275,7 +20285,7 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "7d177203338554097955ce32940d13d2c926ce8450c69998d2a4a5a653bd28ee", + "prepared_review_unit_hash": "2e8fae2a130039a44e0e964c6e9ac4e358b2e3d569655d8d6fe9d95e2d0659ab", "component_budget": { "node_count": 4, "item_count": 4, @@ -20297,58 +20307,55 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Issue 43835 looks unrelated to the three \"huggingface \" reports and should stay separate. Issues 43845, 43846, and 43859 are likely same-title duplicates, but the evidence is thin and mostly title-based; 43845 is the best canonical representative because it is the earliest of that set.", - "confidence": 0.82, - "canonical_issue_reason": "Issue 43845 is the earliest of the likely duplicate \"huggingface \" reports and is the best representative of that same-title cluster.", + "summary": "The cluster looks like weakly related or possibly spammy issue reports, with three nearly identical \"huggingface\" issues and one unrelated \"Fett\" issue. The evidence is too sparse to confidently collapse any pair, so I would not accept any soft-edge merges.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 43845 is the best canonical candidate because it is the earliest of the near-identical \"huggingface\" reports and matches the strongest duplicate signal, but the available metadata is too thin to be fully confident.", "canonical_pr_reason": null, - "best_issue_reason": "43845 is the strongest issue candidate because it is the earliest among the likely same-title duplicates and has no distinguishing detail against the others.", + "best_issue_reason": "43845 is the best representative issue for the near-duplicate subgroup since it is earliest and has the closest match to 43846/43859 by title and metadata. Still, the lack of body content makes this only a tentative choice.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no body-level evidence tying \"Fett\" to the \"huggingface \" reports." + "reason": "The titles are different and there is no body or target evidence tying them to the same underlying issue." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no shared concrete bug description." + "reason": "The only signal is weak generic similarity; there is no concrete evidence they report the same bug." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles and no shared concrete bug description." + "reason": "No meaningful evidence beyond soft similarity, and the titles do not indicate the same problem." }, { "left": "issue:43845", "right": "issue:43846", - "accept": true, - "reason": "Same exact title and similarly sparse bodies; likely the same report despite limited detail." + "accept": false, + "reason": "They are near-identical on title and timing, but the missing descriptions make it impossible to confirm they are the same underlying issue." }, { "left": "issue:43845", "right": "issue:43859", - "accept": true, - "reason": "Same exact title with no distinguishing body evidence; likely the same report, though the evidence is weaker than for the nearer pair." + "accept": false, + "reason": "Same generic title, but no concrete body or target details to prove they are duplicates." }, { "left": "issue:43846", "right": "issue:43859", - "accept": true, - "reason": "Same exact title with no distinguishing body evidence; likely the same report, though the evidence is weaker than for the nearer pair." + "accept": false, + "reason": "The metadata is too sparse to establish that these are the same report rather than separate vague issues." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The 43835 rejections are conservative, but the duplicate claims among 43845/43846/43859 are a bit overconfident. All three have the same title, so a duplicate hypothesis is plausible, but the reasons overstate the evidence by citing 'close creation time' for 43859, which is ~10 hours later than 43845/43846. The summary should stay weaker (e.g. 'likely same-title duplicates') unless there is body-level evidence." - }, - "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -20382,88 +20389,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "2f060cb42bcf4e21a8e0a01d737d46db95a36545011ce8c2fff3672056b10c5d", + "prepared_review_unit_hash": "fd10da3960dbc3b15aff67690cd496a20ffd022cb2f47e157c3f240ecd422fb0", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5785, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5785, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster looks like unrelated low-signal/spam issues rather than true duplicates. A few items share the same placeholder title, but there is no substantive bug or change description tying them together, so all soft duplicate candidates are rejected.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 44418 is the least obviously spammy and the most representative generic issue title in the set, so it is the best available anchor if a canonical item must be chosen.", + "summary": "These are mostly low-information or spam-like issues with only superficial title similarity. No pair has enough concrete evidence to treat as the same underlying problem.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 44418 is the only one that reads like a plausible legitimate ticket ('Documentation Enhancement Suggestion'); the others are generic spam or likely spam.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44418 is the strongest representative of this cluster because it is at least a plausible issue title, unlike the explicit '' entries and the promotional post.", + "best_issue_reason": "44418 is the best issue to keep as the representative because it is the most substantive and specific of the set, even though it is still vague.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Same placeholder title, but no body, targets, or code context; identical spam-like titles are not enough to prove the same underlying issue." + "reason": "Both are titled '', but that is not evidence of the same underlying issue; they are just similarly labeled spam placeholders." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are just '' placeholders with no substantive details, so there is no evidence they report the same bug." + "reason": "Same as above: identical spam titles, but no concrete bug or change described." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is explicit spam and the other is a generic documentation suggestion; they do not describe the same problem." + "reason": "One is spam-like and the other is a vague documentation suggestion; no shared concrete problem." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "Explicit spam vs. a promotional-sounding post; no shared underlying issue is evident." + "reason": "'' and 'Check out \"Google LLC\"' look like unrelated spam/ad content, not the same issue." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "Generic documentation suggestion vs. spam placeholder; no concrete overlap in problem statement." + "reason": "A documentation enhancement request is not the same as a spam placeholder issue." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "These are not the same bug/change; one is a documentation suggestion and the other is spam-like filler." + "reason": "No shared substance beyond both being vague; one is a doc suggestion, the other is spam." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Identical placeholder titles alone are insufficient; there is no body or code-path evidence that they refer to the same underlying issue." + "reason": "Both are titled '', but there is no evidence they describe the same report rather than two separate spam submissions." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder vs. promotional post; no meaningful shared defect or requested change." + "reason": "Spam placeholder versus promotional-looking title; unrelated." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "No substantive overlap beyond being low-signal issues; they do not appear to be the same underlying bug or change." + "reason": "Spam placeholder versus promotional-looking title; unrelated." } ] }, @@ -20493,12 +20500,12 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "08dd66447ed965c6d9303865afc0c1c345ed651cd1b87e6447344392718d6fb0", + "prepared_review_unit_hash": "caab7c9893099df5958def6996728a95f1d7e5877631e18d9fb920580001910e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1646, + "serialized_chars": 1645, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -20506,7 +20513,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1646, + "serialized_chars": 1645, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -20515,27 +20522,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues have the same title and very similar bodies, so they appear to be duplicate reports of the same topic. Issue #45145 is the better canonical choice because it already has an inbound reference.", - "confidence": 0.97, - "canonical_issue_reason": "#45145 is the stronger canonical issue because it has the same title/body topic as #45151 and already has an inbound reference, suggesting it is the one others point to.", + "summary": "#45145 and #45151 look like likely duplicates based on the exact same title, close body lengths, and the soft similarity signal. The metadata does not prove identical bodies, so the assessment is kept conservative. #45145 is the better canonical issue because it has an inbound reference and is slightly more established.", + "confidence": 0.86, + "canonical_issue_reason": "#45145 is the better canonical issue because it has the exact same title as #45151 and already has an inbound reference, giving it a slight edge as the anchor record. The available metadata suggests duplication, but does not let us verify identical bodies.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the best issue candidate for the cluster: it is effectively the same report as #45151 and has slightly more linkage/visibility.", + "best_issue_reason": "#45145 is the best issue to keep: it matches #45151 exactly on title, has a nearby body length, and already has an inbound reference. That makes it the more suitable surviving record for deduplication.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": true, - "reason": "Same title and highly similar issue content; these look like duplicate reports of the same underlying request/topic." + "reason": "Accept as likely duplicate: the titles match exactly and the soft similarity signal supports the same underlying report. The evidence is limited to metadata, so this is a cautious acceptance rather than a claim of identical bodies." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The conclusion is grounded in the packet: the two issues share the exact same title, have close body lengths, and the packet marks them with soft similarity evidence. Choosing #45145 as canonical because it has an inbound reference is also supported. The soft-edge acceptance is reasonable and conservative enough here." + "accept": false, + "feedback": "The conclusion is directionally plausible, but it overstates the evidence. The packet shows identical titles, similar body lengths, and a soft similarity signal, but not that the bodies are the same or that the issues are \"near-identical.\" Also, \"same title/body\" in the canonical_issue_reason is unsupported because the body text is not provided. Please keep the verdict more conservative and grounded in the available metadata." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -20613,12 +20620,12 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "2b26e89bc7d4079aedef045ca67b6866847782267f6043561a5072fc507efbf8", + "prepared_review_unit_hash": "2f2b407273cff70093aed772b7642e21a24c582ad0bc47193905469652fc0c3d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1475, + "serialized_chars": 1476, "estimated_input_tokens": 369, "estimated_eval_tokens": 994 }, @@ -20626,7 +20633,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1475, + "serialized_chars": 1476, "estimated_input_tokens": 369, "estimated_eval_tokens": 994 }, @@ -20635,18 +20642,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two closed issues look like unrelated, content-free entries (personal names plus dates) rather than duplicates of the same bug or request. There is no substantive evidence to cluster them together.", + "summary": "The two closed issues are just short name/date strings and do not describe the same bug or request. They are not duplicates or a meaningful cluster match.", "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: both issues are unrelated and lack any actionable technical content, body text, or links.", + "canonical_issue_reason": "Issue #45762 is marginally more complete as a standalone entry, but neither issue contains actionable problem description.", "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a good global representative; both are effectively empty and appear unrelated.", + "best_issue_reason": "Issue #45762 is the least ambiguous of the two, though both are essentially non-technical placeholder-style issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Different names/dates with no body, filenames, targets, or discussion evidence; they do not appear to describe the same underlying problem." + "reason": "The titles are different personal names with different dates and no shared issue context, code path, or explicit linkage; this does not look like the same underlying bug/request." } ] }, @@ -20676,7 +20683,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "faffbc0fba5303c8715480b0cfd0719e6e31923bdf4f48d9fa707ed87b5f34d0", + "prepared_review_unit_hash": "d2eedcf3f6d0d9123becb11782b6bc8500a32e97c52b3bf2a50d07e5f89ceca4", "component_budget": { "node_count": 2, "item_count": 2, @@ -20698,18 +20705,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are both empty, closed, and appear to be unrelated personal-name/date entries rather than duplicate bug reports. There is no substantive evidence that they describe the same problem.", + "summary": "Both items look like unrelated low-content issues containing personal names/dates, not duplicate bug reports or change requests. The soft similarity is likely due to shared surname only.", "confidence": 0.98, - "canonical_issue_reason": "Neither issue has any body, references, or technical context; if one must be chosen, #45764 is marginally earlier, but neither is a meaningful canonical duplicate target.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No issue in this cluster is a good canonical because both are content-free and do not describe a reproducible bug or request.", + "best_issue_reason": "Neither issue is a good canonical choice: both are empty, non-technical, and appear unrelated apart from the shared surname.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "Only superficial similarity is present (names/dates in titles); there is no indication they refer to the same underlying issue." + "reason": "Different people/names and dates; no shared bug, request, or code-path context, so this is not a duplicate cluster." } ] }, @@ -20739,7 +20746,7 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "b67a7a374b9ffdd96acb5e92842760b887c877ab1fd095b4be16c7f55b81ec40", + "prepared_review_unit_hash": "bcaf14357401832679d17c8feee8437dfa5ff8ea6d93847ef44eca4f42f0a906", "component_budget": { "node_count": 2, "item_count": 2, @@ -20761,18 +20768,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The pair is a false positive: one issue requests a new Qwen token-classification/value-model class, while the other asks for FP8 training support. They concern different features and code paths.", - "confidence": 0.98, - "canonical_issue_reason": "Issue 45810 is the more concrete, self-contained request and the better representative item in this pair.", + "summary": "These are two separate feature requests: one for adding Qwen3_5 token classification/value-model support, and one for FP8 training support. They do not describe the same underlying bug or change.", + "confidence": 0.97, + "canonical_issue_reason": "Not a duplicate cluster; if one must serve as the representative issue, #46295 is broader and already has more inbound attention.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45810 is specific and actionable; it names an explicit model class and use case, unlike the broader FP8 training request.", + "best_issue_reason": "#46295 is the better standalone representative because it has a clearer, broader training-focused request and more inbound references.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Not the same underlying bug or feature: token classification/value-model support vs FP8 training support. No shared targets, filenames, or code-path evidence." + "reason": "Different feature areas and code paths: model head/value-model support vs FP8 training support. They are not the same underlying request." } ] }, @@ -20812,12 +20819,12 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "73c1f85821afe7192a7dcdb304a2cd93bf018c6f859ae27047501f30c7b37b40", + "prepared_review_unit_hash": "2ae9b1dd413645e4c41db0b3c1423eafab2e0d5b163f327cedeba9a1926e3edc", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5725, + "serialized_chars": 5728, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -20825,7 +20832,7 @@ "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5725, + "serialized_chars": 5728, "estimated_input_tokens": 1432, "estimated_eval_tokens": 3120 }, @@ -20834,66 +20841,66 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "There are two clear exact-duplicate issue pairs: 45870/45871 and 45872/45873. The remaining issue 45869 is a different dataset-promo topic, and the cross-pair finance/DataForge matches are only superficial similarity, not the same underlying request.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "The set splits into two exact duplicate issue pairs: 45870\u219445871 and 45872\u219445873. Issue 45869 is a different dataset/sample promotion topic and does not match either pair.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue across the full set; the items form two separate duplicate groups, with 45870/45871 and 45872/45873.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "If a representative issue must be chosen, 45870 is a reasonable anchor because it is an exact duplicate of 45871 and is the earliest of that duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Different topics: a dataset bundle promo versus an example/tutorial about financial time series prediction." + "reason": "Different underlying issue: dataset bundle/sample listing versus an example about financial time series prediction; only broad finance overlap." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Same as 45869/45870; only broad finance wording overlaps, not the same issue." + "reason": "Different underlying issue: dataset bundle/sample listing versus an example about financial time series prediction; only broad finance overlap." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Same title and effectively the same body; this looks like an exact duplicate issue pair." + "reason": "Exact duplicate titles and matching body size; these look like the same example issue." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Both mention financial time series/DataForge, but one is an example issue and the other is about a scraping service; not the same concrete request." + "reason": "Different scope: one is a prediction example, the other is a financial time series plus scraping service; not the same underlying request." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Superficially related by theme, but the artifacts describe different content and should not be merged." + "reason": "Different scope: one is a prediction example, the other is a financial time series plus scraping service; not the same underlying request." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Same mismatch as 45870/45872; shared finance keywords are not enough." + "reason": "Different scope: one is a prediction example, the other is a financial time series plus scraping service; not the same underlying request." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Different underlying ask; not the same bug/change." + "reason": "Different scope: one is a prediction example, the other is a financial time series plus scraping service; not the same underlying request." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Same title and body shape; this is an exact duplicate pair." + "reason": "Exact duplicate titles and matching body size; these look like the same service-related issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: 45870/45871 and 45872/45873 have identical titles and matching body lengths, so calling them duplicates is reasonable, while 45869 is clearly a different dataset-promo topic. The non-duplicate pair verdicts are also appropriately conservative and do not overclaim beyond the visible evidence." + "feedback": "The summary is grounded in the packet: 45870/45871 and 45872/45873 have identical titles and matching body lengths, while 45869 is clearly a different dataset-promotion issue. The soft-edge verdicts are conservative and do not overstate evidence beyond title/body-size duplication." }, "evaluator_used": true, "retried": false, @@ -20920,46 +20927,49 @@ "soft_pairs": [ "issue:46439|issue:46441" ], - "prepared_review_unit_hash": "e4e34aeac53d1047d98300682262f5f0be0b15872c9ed324ae1901311184b226", + "prepared_review_unit_hash": "3963b456205e92776fb4e21cf71cbd9b50129937c9c4df09037c260dc84a2cc8", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1637, - "estimated_input_tokens": 410, - "estimated_eval_tokens": 1076 + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1637, - "estimated_input_tokens": 410, - "estimated_eval_tokens": 1076 + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both issues involve `cache_implementation=\"static\"` and linear-attention/Qwen3.5, but they describe different failures: one is an `early_initialization` shape corruption leading to a runtime shape error, the other is a `KeyError: 'linear_attention'` on Qwen3.5. Similar area, but not enough evidence they are the same underlying bug.", - "confidence": 0.89, - "canonical_issue_reason": null, + "summary": "The two issues are related in theme (static cache + linear-attention/Qwen3.5), but the evidence is too thin to treat them as the same bug. The observed failures differ (`RuntimeError` from shape corruption vs. `KeyError: 'linear_attention'`), and there are no shared files, targets, or explicit cross-links.", + "confidence": 0.44, + "canonical_issue_reason": "Issue 46439 is the better umbrella report because it describes the suspected mechanism (`early_initialization` corrupting linear-attention layers) more generally, which makes it a stronger anchor for related reports.", "canonical_pr_reason": null, - "best_issue_reason": "No clear duplicate canonical: these look like adjacent static-cache failures in the same subsystem, but the reported exceptions and failure modes differ.", + "best_issue_reason": "Issue 46439 is the best representative for the cluster because it is the more root-cause-oriented writeup, but it should be treated as related-not-duplicate evidence rather than a confirmed duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46439", "right": "issue:46441", "accept": false, - "reason": "Both concern static cache and linear attention, but one is a shape-corruption/runtime-shape error during early initialization while the other is a `KeyError` for `linear_attention` in Qwen3.5; they do not clearly describe the same concrete bug." + "reason": "Only soft topical similarity is present. The bugs surface differently (`RuntimeError` vs. `KeyError: 'linear_attention'`), and there are no shared targets/files or explicit links, so this is not enough to conclude the same underlying issue." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "evaluator_result": { + "accept": false, + "feedback": "The analysis overstates the evidence. The packet only supports a soft similarity between two issue titles; there are no shared files/targets or other concrete links. While both mention static cache and linear-attention/Qwen3.5, the specific failures differ (`RuntimeError` from wrong shapes vs. `KeyError: 'linear_attention'`), so treating them as the same bug is not sufficiently grounded. A more conservative conclusion would be that they may be related, not that they are likely the same issue." + }, + "evaluator_used": true, + "retried": true, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -20983,12 +20993,12 @@ "soft_pairs": [ "issue:46551|issue:46552" ], - "prepared_review_unit_hash": "63020bee6bdad8e293203edb593e0d68acee2625aadca172fe5466e0e0ab4c88", + "prepared_review_unit_hash": "6b2b5e1f2591b55c58df5c108bf5b7ebe9168baf44d264ba31ccc8fa7c7afa36", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1706, + "serialized_chars": 1707, "estimated_input_tokens": 427, "estimated_eval_tokens": 1110 }, @@ -20996,7 +21006,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1706, + "serialized_chars": 1707, "estimated_input_tokens": 427, "estimated_eval_tokens": 1110 }, @@ -21005,27 +21015,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues appear to be duplicates based on identical titles, equal body length, and a strong soft-similarity signal. The open issue is the better surviving tracker.", - "confidence": 0.92, - "canonical_issue_reason": "Issue #46552 is the better canonical issue because it is still open and already has inbound references, making it the more useful place to track the bug.", + "summary": "Both issues report the same Wav2Vec2CTCTokenizer v5 bug verbatim; they are duplicates of one another. The open issue is the better canonical record.", + "confidence": 0.99, + "canonical_issue_reason": "issue:46552 is the open copy of the same report and already has inbound references, so it is the better canonical issue to keep.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #46552 is the best issue overall because it matches the same reported problem and is the open ticket most suitable for consolidation.", + "best_issue_reason": "issue:46552 is the strongest single issue candidate because it is open, identical in content to the other issue, and appears to be the active tracking entry.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46551", "right": "issue:46552", "accept": true, - "reason": "Identical titles, equal body length, and the provided soft-similarity evidence support these as the same issue. The packet does not include the bodies themselves, so the verdict is based only on the available signals." + "reason": "The titles and bodies are identical and describe the same tokenizer bug; this is a straightforward duplicate pair." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The conclusion is plausible, but the writeup overstates the evidence. The packet shows identical titles, equal body length, and a soft-similarity signal, but it does not provide the issue bodies themselves, so saying the body is identical is not grounded here. Please keep the duplicate verdict more conservative and base it only on the evidence provided." + "accept": true, + "feedback": "Grounded: the packet shows identical titles, identical body lengths, and maximum soft similarity between the two issues. Choosing the open issue with inbound references as the canonical record is conservative and supported. No evidence is overstated." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -21049,7 +21059,7 @@ "soft_pairs": [ "issue:46657|issue:46692" ], - "prepared_review_unit_hash": "990ba39420ca4d8c71453f2aecee17f8980ae58ab5525af3f9feaff26fa457ee", + "prepared_review_unit_hash": "a0d1801e51db6111b3ed4402de29422eddbe7a0529c34123232eb740da173991", "component_budget": { "node_count": 2, "item_count": 2, @@ -21071,18 +21081,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues describe different failure modes in training: one is a tensor shape mismatch during object detection training, the other is an out-of-memory problem after several epochs in RT-DETR v2. They do not look like duplicates.", - "confidence": 0.96, + "summary": "These two issues are not duplicates: one reports a tensor-size mismatch during object-detection training, while the other reports out-of-memory after several epochs in RT-DETR v2. They may both occur in training, but they describe different failure modes and likely different root causes.", + "confidence": 0.94, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue 46657 is the better representative only because it describes a concrete, actionable training error; however, the two issues are not the same bug, so there is no true canonical duplicate here.", + "best_issue_reason": "46657 is the clearer, more concrete bug report, but it is not the same underlying problem as 46692.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46657", "right": "issue:46692", "accept": false, - "reason": "Different underlying symptoms and likely different root causes: tensor size mismatch vs memory exhaustion. They share a training context, but not the same concrete bug." + "reason": "Reject: the reports describe different issues (tensor shape mismatch vs. OOM after epochs) and do not appear to be the same underlying bug." } ] }, @@ -21112,7 +21122,7 @@ "soft_pairs": [ "issue:46822|issue:46823" ], - "prepared_review_unit_hash": "7e6633f17fd2d39078f68984ac61a268a5958cbbf1bb77e98aeca5cfee57629f", + "prepared_review_unit_hash": "9113889454cfc19309c77b3c85df1a7925c04340d5d3f949d9ffbcd30139bda1", "component_budget": { "node_count": 2, "item_count": 2, @@ -21134,18 +21144,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one is a ZeRO-3/zero.Init partitioning problem causing OOM during multi-GPU model load, while the other is a 4-bit + device_map/max_memory routing bug that incorrectly sends modules to CPU and raises ValueError. They affect different code paths and should stay separate.", + "summary": "These are two separate bug reports in different parts of the model-loading stack: one about ZeRO-3 zero.Init failing to shard a composite submodule and causing OOM, the other about 4-bit auto device mapping with max_memory incorrectly sending modules to CPU and raising a ValueError. They should not be clustered as duplicates.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "No canonical issue: the two issues describe different failure modes and different code paths, so neither subsumes the other.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 46823 is the clearer self-contained report if a single representative is needed, but it is still not a duplicate of 46822.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46822", "right": "issue:46823", "accept": false, - "reason": "Different subsystems and failure modes: ZeRO-3 composite submodule partitioning vs 4-bit device_map/max_memory dispatch to CPU. No evidence they stem from the same underlying bug." + "reason": "Different underlying bugs: ZeRO-3 partitioning/OOM versus 4-bit device_map CPU offload/ValueError. Similar loading context only, not the same fix or code path." } ] }, @@ -21178,7 +21188,7 @@ "issue:46906|issue:46948", "issue:46940|issue:46948" ], - "prepared_review_unit_hash": "2cf29596efbc2e49b0cede2f4c66b337b96c4ed1a49321d83a289e18064826b0", + "prepared_review_unit_hash": "bbf0f291533d1e7cc643658ee1e1c66a62987c5b6ff8089924447d017498c975", "component_budget": { "node_count": 3, "item_count": 3, @@ -21200,30 +21210,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three near-identical daily triage issues for Serge integration failures, but the differing dates strongly suggest separate reports rather than duplicates. I would keep them distinct and not cluster-merge them.", - "confidence": 0.93, - "canonical_issue_reason": "issue:46906 is the best representative because it is the earliest in the set and has the strongest discussion/inbound reference activity.", + "summary": "Three near-identical \"[serge] integration failure triage\" issues for consecutive dates look like recurring daily triage tickets, not the same underlying bug report. I would keep them separate and avoid duplicate-merging.", + "confidence": 0.87, + "canonical_issue_reason": "issue:46906 is the best anchor: it is the earliest of the three and has the strongest inbound reference count, so it is the most representative triage ticket for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "It is the most visible and well-referenced of the three, making it the most useful cluster representative if one must be chosen.", + "best_issue_reason": "issue:46906 has the most references and earliest timestamp, making it the strongest single issue to point cluster readers to.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46906", "right": "issue:46940", "accept": false, - "reason": "Same template/title pattern, but the date change indicates a separate daily triage issue rather than the same underlying report." + "reason": "Same template/title, but the date-specific triage context suggests separate daily failure reports rather than one shared underlying issue." }, { "left": "issue:46906", "right": "issue:46948", "accept": false, - "reason": "These appear to be consecutive daily triage issues; similarity is from the standardized title, not shared bug identity." + "reason": "These appear to be consecutive-day triage entries; similarity is high, but there is no evidence they describe the exact same bug or fix." }, { "left": "issue:46940", "right": "issue:46948", "accept": false, - "reason": "Both are date-stamped triage entries for different days, so they should not be treated as duplicates." + "reason": "Likely separate daily integration-failure tracking issues, not duplicate reports of one concrete problem." } ] }, @@ -21253,12 +21263,12 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "eb69da3ec10d765a56fa420c0d0f9d0fa9cfe18a9710ba85f97a8f730bf68861", + "prepared_review_unit_hash": "1b7f64e9b178bcfb65c1f8832bd58dc865040cef7de327a31a0c5686ce9d60c6", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7224, + "serialized_chars": 7223, "estimated_input_tokens": 1806, "estimated_eval_tokens": 3868 }, @@ -21266,7 +21276,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7224, + "serialized_chars": 7223, "estimated_input_tokens": 1806, "estimated_eval_tokens": 3868 }, @@ -21275,18 +21285,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are VibeVoice-related, but they are different changes: #40546 implements the base VibeVoice model, while #42781 adds VibeVoice Realtime plus acoustic tokenizer support and other follow-on wiring. Shared filenames alone aren\u2019t enough to treat them as duplicates.", - "confidence": 0.91, + "summary": "Both PRs target VibeVoice, but they are not the same concrete change: #40546 is the initial VibeVoice implementation, while #42781 is a draft for VibeVoice Realtime with additional tokenizer/processing and generation changes. Shared files alone are not enough to treat them as duplicates.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "#42781 is the broader, later follow-on PR and is the better representative of the VibeVoice work, but it is not a duplicate of #40546.", + "canonical_pr_reason": "#40546 is the broader, primary implementation PR with substantial review activity and the initial model integration; #42781 builds on the same model family but is a separate realtime feature branch.", "best_issue_reason": null, - "best_pr_reason": "#42781 is the best single PR to represent the cluster because it contains the more complete VibeVoice integration and extends the initial implementation rather than repeating it.", + "best_pr_reason": "#40546 is the best representative of the base VibeVoice integration because it is the main non-draft implementation PR and covers the foundational code path.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Reject: #40546 is the base VibeVoice implementation, while #42781 adds realtime behavior and acoustic tokenizer/model wiring. They overlap on shared model files but do not appear to be the same concrete change." + "reason": "They share the VibeVoice subsystem, but #42781 is a distinct realtime add-on with extra generation/processing and acoustic tokenizer work, not the same bug or change as the initial implementation in #40546." } ] }, @@ -21316,12 +21326,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "05703f1aba5d4d50eb012da1742c6659f571adc532bd86c3faac4b2f673cad85", + "prepared_review_unit_hash": "5892007c7a6f96cc7c7b2f5efc12cd0ef7ddb11d0c91b3a6c8c89e476eabcdf7", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4676, + "serialized_chars": 4674, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -21329,7 +21339,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4676, + "serialized_chars": 4674, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -21338,18 +21348,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are unrelated feature additions in different model families (TimesFM covariates vs. DINOv3 image classification), so they are not duplicates.", - "confidence": 0.99, + "summary": "These PRs are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. The soft similarity is superficial (both add docs/tests and model support), not the same underlying change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the two pull requests touch different models, different files, and different code paths, so they cannot represent the same underlying change.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "No global best PR for this cluster: 40755 is a TimesFM forecasting enhancement, while 41224 is a DINOv3 classification feature; neither can stand in for the other.", + "best_pr_reason": "PR 41224 is the stronger representative artifact: it is a focused, concrete model capability addition with clear implementation scope, docs, and tests.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different features for different models: TimesFM covariate forecasting vs. DINOv3 image classification. No shared bug or change path, and they would not be merged into one PR." + "reason": "Different models, different tasks, and different code paths; they do not fix the same bug or implement the same change." } ] }, @@ -21470,154 +21480,148 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "29591647ba533a1a2ce8d99bb587a5762fce8bd9221ae2a27620a5b59876cb04", + "prepared_review_unit_hash": "bec53376d46f663f496be2797675ca0d4641f87f631704f342cd2a645d80baa8", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 18, - "serialized_chars": 34786, - "estimated_input_tokens": 8697, - "estimated_eval_tokens": 17650 + "serialized_chars": 34780, + "estimated_input_tokens": 8695, + "estimated_eval_tokens": 17646 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 18, - "serialized_chars": 34786, - "estimated_input_tokens": 8697, - "estimated_eval_tokens": 17650 + "serialized_chars": 34780, + "estimated_input_tokens": 8695, + "estimated_eval_tokens": 17646 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Mostly unrelated pull requests were grouped by superficial similarity. The only true duplicate/change overlap is the pair of `fixed_cross_entropy` kwargs PRs; everything else targets different subsystems or different features.", - "confidence": 0.97, + "summary": "This cluster is mostly unrelated PRs across different subsystems. The only true near-duplicate is the pair adding kwargs support to `fixed_cross_entropy` in `loss_utils.py` (43251/43254); all other pairs are just thematic or subsystem similarity, not the same underlying change.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the best canonical representative for the duplicate pair: it has the explicit issue link, more discussion/review activity, and the same `cross_entropy` fix as 43254.", + "canonical_pr_reason": "PR 43251 is the best canonical PR in the only duplicate pair: it targets the same `fixed_cross_entropy` change, explicitly links the tracked issue, and presents the more complete implementation of forwarding supported kwargs to `nn.functional.cross_entropy`.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the strongest overall match in the cluster because it is the only accepted duplicate pair member and the more complete implementation of that change.", + "best_pr_reason": "PR 43251 is the strongest global representative here because it is the clearest duplicate-center candidate and the only PR that cleanly matches another submission of the same fix in the same code path.", "soft_edge_verdicts": [ { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Mistral tokenizer missing-methods work vs. MobileLLM model addition; different code paths and goals." + "reason": "Different code paths and goals: Mistral tokenizer methods vs adding a new MobileLLM model. Shared \u2018model support\u2019 theme is too broad." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer API addition vs. Executorch dynamic-shape export test; unrelated changes." + "reason": "Tokenizer API addition and ExecuTorch dynamic-shape test are unrelated changes; they would not be merged as one PR." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor/auto-loading robustness work vs. new MobileLLM model support; not the same change." + "reason": "Both touch pretrained loading/processor-related code, but one is a robustness fix for processor auto-loading and the other is a new model addition." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Cross-entropy kwargs fix vs. MobileLLM model addition; unrelated." + "reason": "Loss-function kwargs support and MobileLLM model addition are completely different features." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Loss utility kwargs fix vs. Mistral4 model/test changes; different bug and subsystem." + "reason": "Different bug fixes in different model families; no shared concrete code-path problem." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Executorch export test vs. Mistral4 test/model work; no shared underlying change." + "reason": "ExecuTorch export tests and Mistral4 test fixes are unrelated test additions, not the same underlying change." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Dynamic-shape export test vs. model addition; unrelated." + "reason": "Dynamic-shape export testing is unrelated to adding MobileLLM." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "New MobileLLM model support vs. config resolution fix; different problem." + "reason": "New model implementation vs class/config resolution ordering fix; no duplicate bug or mergeable shared fix." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "MobileLLM model addition vs. Falcon output-collection refactor; unrelated." - }, - { - "left": "pull_request:43823", - "right": "pull_request:45213", - "accept": false, - "reason": "Model implementation PR vs. meta skill/docs/tooling PR; not the same artifact or bug." + "reason": "MobileLLM model addition and Falcon output-collection refactor are different model changes with different purposes." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon model refactor vs. model-creation skill/meta changes; unrelated." + "reason": "Falcon refactor and model-creation skill are not the same change; the latter is a meta/tooling PR, not a code-path fix." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output-interface refactor vs. Mistral4 tests/model changes; different code paths." + "reason": "Falcon refactor and Mistral4 test fixes touch different models and different problems." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test/model work vs. CLIP-like `layrnorm` spelling fix across conversion/model files; unrelated." + "reason": "Mistral4 test fixes and CLIP-family `layrnorm` rename are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:46267", "accept": false, - "reason": "Mistral4 tests/model changes vs. chat parsing/serve overhaul; completely different functionality." + "reason": "Mistral4 test fixes and chat parsing updates are different subsystems and changes." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 model/test updates vs. TDT loss kernel integration; different feature area." + "reason": "Mistral4 tests and TDT loss kernel integration are unrelated changes." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 test/model changes vs. config resolution fix; unrelated." + "reason": "Mistral4 test fixes and `cls.config` resolution ordering are different problems with no shared code path." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Concrete model/test changes vs. a model-creation skill draft; not the same change." + "reason": "Shared auto/modeling files are incidental; one is Mistral4 tests and the other is a model-creation skill draft." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM implementation vs. model-creation skill/meta tooling; unrelated." + "reason": "New model implementation and the model-creation skill are not duplicate code changes." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs modify `fixed_cross_entropy` to forward supported `cross_entropy` kwargs for the same issue and same function, so they are duplicate implementations of the same fix." + "reason": "Same function, same file, same issue target, and both add support for passing cross-entropy kwargs; they are clearly alternative implementations of the same fix and could be merged into one PR." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative on the soft-edge pairs. The only accepted pair, 43251/43254, clearly targets the same `fixed_cross_entropy` kwargs bug in the same file and is plausibly mergeable into one bugfix PR. The rest of the pairwise rejections are consistent with the diffs and do not overstate evidence." + "feedback": "The summary is conservative and grounded in the packet. The only clearly mergeable duplicate pair is PR 43251 / 43254: they touch the same file, same function, and share the same issue target with overlapping `cross_entropy` kwargs support. The other pairwise rejects stay within the evidence shown (mostly only soft similarity with no shared file or issue target), and no overclaiming stands out." }, "evaluator_used": true, "retried": false, @@ -21644,7 +21648,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "b76197698319c9dafb63bb65acc85e9e763ceca7b9e22d2ee4558cffa6f6b10c", + "prepared_review_unit_hash": "4b5c38ac0656831c6f2c800a4c003ba2fef06a93eb2223536d9f402f92e39b29", "component_budget": { "node_count": 2, "item_count": 2, @@ -21666,18 +21670,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two broad workflow-maintenance PRs overlap on several GitHub Actions files, but they target different upgrade sets and different concerns: one is for Node 24 compatibility, the other updates Docker build actions to newer releases. They are related in theme but not the same underlying change.", - "confidence": 0.94, + "summary": "Both pull requests are broad dependency bumps in GitHub Actions, but they target different action families and different compatibility goals, so they are not duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the cleaner canonical representative because its goal is explicit and cohesive: upgrading workflow actions for Node 24 compatibility across many files.", + "canonical_pr_reason": "No single canonical PR for duplication here: PR 42975 refreshes multiple GitHub Actions to Node 24-compatible versions, while PR 42976 specifically upgrades docker/build-push-action across Docker-related workflows.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the best standalone PR for this cluster because it has the clearest, most specific purpose and broadest coherent action-version refresh; 42976 is a separate Docker-action upgrade.", + "best_pr_reason": "PR 42975 is the broader workflow-action maintenance PR and is the closest representative of the cluster, but it still should not subsume PR 42976.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Do not merge: both are workflow dependency bumps, but they change different action families and solve different upgrade targets. Shared filenames only show overlapping CI surface, not the same concrete bug or one mergeable change set." + "reason": "Different underlying changes: 42975 updates several GitHub Actions for Node 24 compatibility, while 42976 only bumps docker/build-push-action in Docker image workflows. They touch overlapping files but not the same concrete bug/change." } ] }, @@ -21712,12 +21716,12 @@ "pull_request:43448|pull_request:45490", "pull_request:45490|pull_request:46562" ], - "prepared_review_unit_hash": "7c265557ed8ecd816815e7e510ac760354484b2843d47463ac2940cf70807c7d", + "prepared_review_unit_hash": "a7b8f1f88d5bf84d9c5d891b422dca03090c812286368a17c2e4d2fffacb601e", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 13046, + "serialized_chars": 13047, "estimated_input_tokens": 3262, "estimated_eval_tokens": 6780 }, @@ -21725,7 +21729,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 13046, + "serialized_chars": 13047, "estimated_input_tokens": 3262, "estimated_eval_tokens": 6780 }, @@ -21734,36 +21738,36 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All four PRs are distinct model-addition changes (ViT NEPA, Molmo, CTSM, Step3p7). They share generic registry/docs boilerplate in auto mappings and the toctree, but they target different model directories and code paths, so none should be merged as duplicates.", - "confidence": 0.98, + "summary": "These are four separate model-addition PRs (ViT NEPA, Molmo, CTSM, Step3p7). They share only broad Hugging Face model-integration boilerplate such as docs and auto-registry updates, not the same concrete code change or bug.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR: the items are unrelated model integrations rather than duplicates of one another.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "#46562 is the strongest standalone representative by review activity and scope, but it is still just the Step3p7 model-addition PR, not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different model integrations: ViT NEPA vs Molmo. Shared files are only common docs/auto-registry plumbing, not the same underlying bug or change." + "reason": "Different models and different integrations; shared auto/docs files are generic boilerplate, not the same underlying change." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "These add different models (ViT NEPA and CTSM) with different model-specific files and behaviors; overlap is limited to shared registration/docs." + "reason": "ViT NEPA vs CTSM are unrelated model additions; overlap is limited to registry/docs plumbing." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo and CTSM are unrelated model additions. Similar filenames reflect broad framework wiring, not a common concrete code-path fix." + "reason": "Molmo and CTSM are distinct model implementations; similarity is only in common framework registration and docs updates." }, { "left": "pull_request:45490", "right": "pull_request:46562", "accept": false, - "reason": "CTSM and Step3p7 are separate model implementations. Despite overlapping auto-mapping/docs changes, they are not the same change and could not be merged into one PR." + "reason": "CTSM time-series support and Step3p7 multimodal support are different concrete model changes; shared auto-mapping/docs edits are not enough to treat them as duplicates." } ] }, @@ -21812,130 +21816,130 @@ "pull_request:46539|pull_request:46922", "pull_request:46886|pull_request:46922" ], - "prepared_review_unit_hash": "33fe807409d19d9a66b6abc1ced6e4cb1eebe9d7385c764e9286e9188a4fbd97", + "prepared_review_unit_hash": "f8794db6705d784f5f547f4244f1bcebe0745b7dace2bf6a24e36d4f31ccb5fe", "component_budget": { "node_count": 7, "item_count": 7, "soft_pair_count": 15, - "serialized_chars": 19599, - "estimated_input_tokens": 4900, - "estimated_eval_tokens": 10056 + "serialized_chars": 19604, + "estimated_input_tokens": 4901, + "estimated_eval_tokens": 10058 }, "budget": { "node_count": 7, "item_count": 7, "soft_pair_count": 15, - "serialized_chars": 19599, - "estimated_input_tokens": 4900, - "estimated_eval_tokens": 10056 + "serialized_chars": 19604, + "estimated_input_tokens": 4901, + "estimated_eval_tokens": 10058 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly unrelated open PRs that happen to share superficial similarity. The only true duplicate-like pair is the two YOLOS fixes, both tied to the same device-mismatch issue; the rest should stay separate.", + "summary": "Mostly unrelated PRs; the only true duplicate pair is the two YOLOS fixes, where PR #46922 subsumes the device-placement correction and the no-split-module guard.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46886 is the better canonical representative: it has the clearest title, an explicit issue target (46869), and the patch directly addresses the YOLOS device_map mismatch in the model code.", + "canonical_pr_reason": "PR #46922 is the best canonical PR because it directly fixes the YOLOS device mismatch in the forward path and also adds the supporting no-split module safeguard for the same bug.", "best_issue_reason": null, - "best_pr_reason": "PR #46886 is the strongest overall choice because it is the most specific and complete fix for the shared YOLOS bug, while #46922 looks like an alternate/partial variant on the same issue.", + "best_pr_reason": "PR #46922 is the strongest representative of the cluster: it targets the concrete YOLOS device_map=\"auto\" failure and appears to contain the more complete fix.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Different models and different changes; both are test-related but not the same bug or mergeable fix." + "reason": "Both are test-only changes in different model areas; they do not describe the same underlying bug or mergeable fix." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper test/tokenization work is unrelated to the FP8 dequantization fix." + "reason": "Whisper test cleanup vs FP8 dequant/loading fix are unrelated code paths." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Whisper decoding/tests and generation/watermarking validation are unrelated changes." + "reason": "Whisper tokenizer tests and generation watermarking/validation changes are unrelated." }, { "left": "pull_request:43291", "right": "pull_request:46922", "accept": false, - "reason": "Whisper tokenizer/test PR and YOLOS model fix do not target the same underlying problem." + "reason": "Whisper test fixes are unrelated to the YOLOS device mismatch fix." }, { "left": "pull_request:43665", "right": "pull_request:46539", "accept": false, - "reason": "Both touch tests, but they fix unrelated issues in different subsystems." + "reason": "These are separate test-only PRs for different models and different failure modes." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "A noisy test-only PR and a model-loading/FP8 dequant PR are not the same change." + "reason": "CLIP/ViT test assertions do not match the FP8 dequantization and loading-path change." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "CLIP/ViT test edits and generation watermarking validation are unrelated." + "reason": "Different subsystems and failure modes: model test cleanup vs generation config validation." }, { "left": "pull_request:43665", "right": "pull_request:46922", "accept": false, - "reason": "Test placeholder edits are unrelated to the YOLOS device-mismatch fix." + "reason": "The test-only CLIP/ViT PR does not fix the YOLOS runtime device mismatch." }, { "left": "pull_request:44815", "right": "pull_request:46539", "accept": false, - "reason": "FP8 dequantization/model loading changes are unrelated to Phimoe CI/test-environment fixes." + "reason": "FP8 quantization/loading fixes are unrelated to the Phimoe CI/integration-test environment changes." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "Different code paths: FP8 quantization handling versus generation config validation." + "reason": "These touch different functionality: dequantization internals vs generation watermarking validation." }, { "left": "pull_request:44815", "right": "pull_request:46922", "accept": false, - "reason": "No shared concrete bug; FP8 dequant work is unrelated to YOLOS device placement." + "reason": "No shared bug: FP8 dequant fix is unrelated to YOLOS device placement." }, { "left": "pull_request:46380", "right": "pull_request:46539", "accept": false, - "reason": "Generation watermarking validation and Phimoe integration-test CI fixes are different problems." + "reason": "Generation-config watermarking validation and Phimoe CI setup are not the same underlying issue." }, { "left": "pull_request:46380", "right": "pull_request:46922", "accept": false, - "reason": "These PRs address unrelated generation/validation versus YOLOS device-map behavior." + "reason": "Generation config validation changes are unrelated to the YOLOS mismatch bug." }, { "left": "pull_request:46539", "right": "pull_request:46922", "accept": false, - "reason": "CI memory/device-map test infrastructure is unrelated to the YOLOS model fix." + "reason": "Phimoe integration-test CI adjustments are unrelated to YOLOS runtime behavior." }, { "left": "pull_request:46886", "right": "pull_request:46922", "accept": true, - "reason": "Both target the same YOLOS device_mismatch with device_map=\"auto\" on the same file and same issue target, and could plausibly be merged into one fix." + "reason": "Both PRs target the same YOLOS device-mismatch problem in the same file and same issue target; #46922 adds the actual tensor device move plus the split-module guard, making them mergeable into one fix." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the only accepted soft edge is conservative: PRs 46886 and 46922 share the same YOLOS file and explicit issue target, so they plausibly represent mergeable fixes for the same bug. The other pairwise rejections are also appropriately conservative because the remaining PRs touch clearly different subsystems and do not look mergeable into one bugfix." + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The only accepted duplicate pair, #46886 and #46922, is well-supported by the shared YOLOS file and shared issue target, and the mergeability claim is plausible because #46922 subsumes the device fix plus the no-split-module safeguard. The other pair rejections are also conservative and match the diffs shown." }, "evaluator_used": true, "retried": false, @@ -21948,13 +21952,13 @@ "component_index": 8, "component_count": 36, "review_unit_index": 1, - "review_unit_count": 1, + "review_unit_count": 2, "status": "reviewed", "reason": null, "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-43451-18", + "cluster_id": "cluster-43451-24", "nodes": [ "pull_request:43451", "pull_request:45144", @@ -21963,464 +21967,588 @@ "pull_request:46658", "pull_request:46707", "pull_request:46836", + "pull_request:46850", "pull_request:46891", "pull_request:46895", + "pull_request:46902", "pull_request:46903", "pull_request:46912", + "pull_request:46920", "pull_request:46935", "pull_request:46953", "pull_request:46954", "pull_request:46957", "pull_request:46958", "pull_request:46961", - "pull_request:46964" + "pull_request:46964", + "pull_request:46965", + "pull_request:46966", + "pull_request:46968" ], "soft_pairs": [ - "pull_request:43451|pull_request:46836", - "pull_request:43451|pull_request:46196", - "pull_request:43451|pull_request:46957", - "pull_request:43451|pull_request:46953", - "pull_request:43451|pull_request:46891", + "pull_request:46895|pull_request:46958", + "pull_request:46903|pull_request:46958", + "pull_request:46902|pull_request:46965", + "pull_request:46902|pull_request:46957", + "pull_request:46954|pull_request:46965", + "pull_request:46954|pull_request:46961", + "pull_request:46180|pull_request:46658", + "pull_request:46658|pull_request:46903", + "pull_request:46891|pull_request:46964", "pull_request:45144|pull_request:46912", + "pull_request:46891|pull_request:46954", + "pull_request:43451|pull_request:46954", + "pull_request:46954|pull_request:46957", + "pull_request:46903|pull_request:46964", + "pull_request:46964|pull_request:46966", + "pull_request:46196|pull_request:46902", + "pull_request:46658|pull_request:46902", + "pull_request:43451|pull_request:46902", + "pull_request:46935|pull_request:46957", + "pull_request:46958|pull_request:46961", "pull_request:45144|pull_request:46196", "pull_request:43451|pull_request:45144", + "pull_request:46891|pull_request:46935", + "pull_request:43451|pull_request:46964", + "pull_request:43451|pull_request:46935", "pull_request:45144|pull_request:46707", + "pull_request:46935|pull_request:46961", "pull_request:45144|pull_request:46658", - "pull_request:46180|pull_request:46658", - "pull_request:46196|pull_request:46957", - "pull_request:46196|pull_request:46836", - "pull_request:46196|pull_request:46953", - "pull_request:46196|pull_request:46891", + "pull_request:46935|pull_request:46953", + "pull_request:46707|pull_request:46920", + "pull_request:46912|pull_request:46920", + "pull_request:46658|pull_request:46964", + "pull_request:46836|pull_request:46964", + "pull_request:46196|pull_request:46964", + "pull_request:46958|pull_request:46965", + "pull_request:43451|pull_request:46920", + "pull_request:46920|pull_request:46957", + "pull_request:46957|pull_request:46958", + "pull_request:46836|pull_request:46920", + "pull_request:46891|pull_request:46958", + "pull_request:43451|pull_request:46836", + "pull_request:46891|pull_request:46957", "pull_request:46658|pull_request:46836", "pull_request:43451|pull_request:46658", - "pull_request:46658|pull_request:46707", - "pull_request:46658|pull_request:46957", - "pull_request:46658|pull_request:46891", + "pull_request:43451|pull_request:46196", + "pull_request:46953|pull_request:46957", "pull_request:46707|pull_request:46953", - "pull_request:43451|pull_request:46707", + "pull_request:43451|pull_request:46891", + "pull_request:46658|pull_request:46850", + "pull_request:46891|pull_request:46953", + "pull_request:46658|pull_request:46707", "pull_request:46707|pull_request:46891", + "pull_request:43451|pull_request:46707", + "pull_request:46836|pull_request:46891", + "pull_request:43451|pull_request:46953", + "pull_request:46836|pull_request:46957", + "pull_request:46658|pull_request:46957", + "pull_request:46961|pull_request:46965", "pull_request:46707|pull_request:46836", + "pull_request:46957|pull_request:46965", + "pull_request:46953|pull_request:46968", "pull_request:46707|pull_request:46957", - "pull_request:46836|pull_request:46957", - "pull_request:46836|pull_request:46891", - "pull_request:46891|pull_request:46957", - "pull_request:46891|pull_request:46953", - "pull_request:46903|pull_request:46958", - "pull_request:46658|pull_request:46903", - "pull_request:46903|pull_request:46964", + "pull_request:46196|pull_request:46957", "pull_request:46912|pull_request:46957", + "pull_request:46957|pull_request:46968", + "pull_request:46850|pull_request:46957", "pull_request:43451|pull_request:46912", - "pull_request:46891|pull_request:46912", - "pull_request:46836|pull_request:46912", - "pull_request:46707|pull_request:46912", - "pull_request:46935|pull_request:46957", - "pull_request:46891|pull_request:46935", - "pull_request:43451|pull_request:46935", - "pull_request:46935|pull_request:46961", - "pull_request:46935|pull_request:46953", - "pull_request:46953|pull_request:46957", - "pull_request:46836|pull_request:46953", - "pull_request:46954|pull_request:46961", - "pull_request:46891|pull_request:46954", - "pull_request:43451|pull_request:46954", - "pull_request:46954|pull_request:46957", - "pull_request:46953|pull_request:46954", - "pull_request:46958|pull_request:46961", - "pull_request:46957|pull_request:46958", - "pull_request:46891|pull_request:46958", - "pull_request:43451|pull_request:46958", - "pull_request:46957|pull_request:46961", - "pull_request:46891|pull_request:46961", - "pull_request:43451|pull_request:46961", - "pull_request:46953|pull_request:46961", - "pull_request:46836|pull_request:46961", - "pull_request:46891|pull_request:46964", - "pull_request:43451|pull_request:46964", - "pull_request:46836|pull_request:46964", - "pull_request:46658|pull_request:46964", - "pull_request:46707|pull_request:46964", - "pull_request:46895|pull_request:46958" + "pull_request:46850|pull_request:46891", + "pull_request:46891|pull_request:46968", + "pull_request:43451|pull_request:46957", + "pull_request:46196|pull_request:46836", + "pull_request:43451|pull_request:46850" ], - "prepared_review_unit_hash": "9dadb764b532c9a64faa6d1b3569b5db22d0e09052c06072d044045546fb3edd", + "prepared_review_unit_hash": "55caed96e7fac2c508287592d92adabf7a36a4c3ba90a9e4541972b679923bbf", "component_budget": { - "node_count": 18, - "item_count": 18, - "soft_pair_count": 64, - "serialized_chars": 68740, - "estimated_input_tokens": 17185, - "estimated_eval_tokens": 34626 + "node_count": 24, + "item_count": 24, + "soft_pair_count": 86, + "serialized_chars": 90472, + "estimated_input_tokens": 22618, + "estimated_eval_tokens": 45492 }, "budget": { - "node_count": 18, - "item_count": 18, - "soft_pair_count": 64, - "serialized_chars": 68740, - "estimated_input_tokens": 17185, - "estimated_eval_tokens": 34626 + "node_count": 24, + "item_count": 24, + "soft_pair_count": 72, + "serialized_chars": 79405, + "estimated_input_tokens": 19852, + "estimated_eval_tokens": 39960 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, - "split": false, + "split": true, "analyst_result": { - "summary": "This cluster is mostly a false-positive grouping of unrelated PRs that share broad patterns like model additions, infrastructure tweaks, or test-only changes. The only clear duplicate-like pair is the two CohereASR training-loss fixes, which target the same bug and code path.", - "confidence": 0.93, + "summary": "This cluster is mostly a set of unrelated PRs that happen to look similar at a text/surface level. The only clear duplicate pair is the CohereASR training-loss double-shift fix (46895 and 46958); the rest are different model additions, distinct bug fixes, CI/docs changes, or separate infrastructure work.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46895 is the clearest representative of the only true duplicate pair: it fixes the CohereASR double-shift training-loss bug directly and has an explicit issue link.", + "canonical_pr_reason": "PR 46895 is the best canonical PR because it is the clearest representative of the one genuine duplicate pair: it fixes the same CohereASR training-loss double-shift bug as 46958, targets the same issue, and has the stronger review history.", "best_issue_reason": null, - "best_pr_reason": "PR #46895 is the best overall representative because it is a concrete, self-contained fix for the shared CohereASR loss-shift bug and is easier to anchor as the canonical duplicate.", + "best_pr_reason": "PR 46895 is the strongest overall representative: it is the only clearly duplicated change in the set and is the more established version of that CohereASR fix.", "soft_edge_verdicts": [ { - "left": "pull_request:43451", - "right": "pull_request:46836", + "left": "pull_request:46895", + "right": "pull_request:46958", + "accept": true, + "reason": "Same CohereASR training-loss double-shift bug in the same code path/files; 46958 is an alternate implementation of the same fix." + }, + { + "left": "pull_request:46903", + "right": "pull_request:46958", "accept": false, - "reason": "Different model additions: Molmo2 vs Unlimited OCR; same docs/auto-mapping pattern is too broad." + "reason": "Both are training-loss fixes, but for different models (PPFormulaNet vs CohereASR); not the same underlying change." }, { - "left": "pull_request:43451", - "right": "pull_request:46196", + "left": "pull_request:46902", + "right": "pull_request:46965", "accept": false, - "reason": "Molmo2 model addition vs wav2vec2 SDPA dispatch; unrelated changes." + "reason": "Offloading/save-state changes vs DeepSeek V4 cache handling; unrelated bug fixes." }, { - "left": "pull_request:43451", + "left": "pull_request:46902", "right": "pull_request:46957", "accept": false, - "reason": "Molmo2 model addition vs generation/FP8 optimization; no shared bug or code path." + "reason": "Different areas: disk-offloaded buffer saving vs generation decode optimization." }, { - "left": "pull_request:43451", - "right": "pull_request:46953", + "left": "pull_request:46954", + "right": "pull_request:46965", "accept": false, - "reason": "Molmo2 model addition vs RMSNorm dtype cleanup; unrelated." + "reason": "Different model internals: FusedRMSNormGated dtype cleanup vs DeepSeek V4 use_cache behavior." }, { - "left": "pull_request:43451", - "right": "pull_request:46891", + "left": "pull_request:46954", + "right": "pull_request:46961", "accept": false, - "reason": "Model addition vs CI workflow tweak; not the same change." + "reason": "A code clean-up in model modules vs a workflow branch-normalization PR; unrelated." }, { - "left": "pull_request:45144", - "right": "pull_request:46912", + "left": "pull_request:46180", + "right": "pull_request:46658", "accept": false, - "reason": "MiMo model addition vs RoPE test expansion; unrelated." + "reason": "Different model additions (Fun-ASR-Nano vs Step-3.7-Flash), not the same change." }, { - "left": "pull_request:45144", - "right": "pull_request:46196", + "left": "pull_request:46658", + "right": "pull_request:46903", "accept": false, - "reason": "MiMo model addition vs wav2vec2 SDPA dispatch; different subsystems." + "reason": "Step-3.7-Flash model build/conversion work vs PPFormulaNet loss fix; unrelated." }, { - "left": "pull_request:43451", - "right": "pull_request:45144", + "left": "pull_request:46891", + "right": "pull_request:46964", "accept": false, - "reason": "Different model additions (Molmo2 vs Xiaomi MiMo-V2); same broad category only." + "reason": "CI/workflow changes vs quantized-PEFT training validation; different purpose and code path." }, { "left": "pull_request:45144", - "right": "pull_request:46707", + "right": "pull_request:46912", "accept": false, - "reason": "MiMo model addition vs native FSDP2 integration; unrelated changes." + "reason": "MiMo-V2 model addition vs RoPE test expansion; not the same underlying bug/change." }, { - "left": "pull_request:45144", - "right": "pull_request:46658", + "left": "pull_request:46891", + "right": "pull_request:46954", "accept": false, - "reason": "Different model additions: MiMo-V2 vs Step-3.7-Flash." + "reason": "CI integration changes vs model dtype clean-up; unrelated." }, { - "left": "pull_request:46180", - "right": "pull_request:46658", + "left": "pull_request:43451", + "right": "pull_request:46954", "accept": false, - "reason": "Fun-ASR-Nano model addition vs Step-3.7-Flash model work; unrelated." + "reason": "Molmo2 model addition vs FusedRMSNormGated dtype change; different work." }, { - "left": "pull_request:46196", + "left": "pull_request:46954", "right": "pull_request:46957", "accept": false, - "reason": "SDPA dispatch changes for audio models vs FP8 generation optimization; different bugs." + "reason": "Model dtype/module cleanup vs generation kernel optimization; unrelated." }, { - "left": "pull_request:46196", - "right": "pull_request:46836", + "left": "pull_request:46903", + "right": "pull_request:46964", "accept": false, - "reason": "wav2vec2 SDPA dispatch vs Unlimited OCR model addition; unrelated." + "reason": "Different fixes in different models; only share a broad training context." }, { - "left": "pull_request:46196", - "right": "pull_request:46953", + "left": "pull_request:46964", + "right": "pull_request:46966", "accept": false, - "reason": "wav2vec2 SDPA dispatch vs removing a default dtype in fused norm modules; unrelated." + "reason": "Quantization validation vs Gemma4Unified sequence-classification support; different feature areas." }, { "left": "pull_request:46196", - "right": "pull_request:46891", + "right": "pull_request:46902", "accept": false, - "reason": "Audio-model SDPA support vs CI workflow changes; not the same change." + "reason": "SDPA attention docs/code vs offloading/save conversion; unrelated." }, { "left": "pull_request:46658", - "right": "pull_request:46836", + "right": "pull_request:46902", "accept": false, - "reason": "Step-3.7-Flash model work vs Unlimited OCR model addition; unrelated." + "reason": "Step-3.7-Flash model work vs offloading/save-state bugfix; unrelated." }, { "left": "pull_request:43451", - "right": "pull_request:46658", + "right": "pull_request:46902", "accept": false, - "reason": "Molmo2 vs Step-3.7-Flash are different model additions." + "reason": "Molmo2 model addition vs offloading/save-state bugfix; unrelated." }, { - "left": "pull_request:46658", - "right": "pull_request:46707", + "left": "pull_request:46935", + "right": "pull_request:46957", "accept": false, - "reason": "Step-3.7-Flash model conversion work vs FSDP2 infrastructure; unrelated." + "reason": "Trackio callback finalization vs generation decode optimization; different subsystems." }, { - "left": "pull_request:46658", - "right": "pull_request:46957", + "left": "pull_request:46958", + "right": "pull_request:46961", "accept": false, - "reason": "Step-3.7-Flash model code vs FP8 decode optimization; no shared concrete bug." + "reason": "CohereASR loss fix vs Serge workflow normalization; unrelated." }, { - "left": "pull_request:46658", - "right": "pull_request:46891", + "left": "pull_request:45144", + "right": "pull_request:46196", "accept": false, - "reason": "Step-3.7-Flash model work vs PEFT CI workflow changes; unrelated." + "reason": "Different model additions/fixes (MiMo-V2 vs wav2vec2 SDPA dispatch)." }, { - "left": "pull_request:46707", - "right": "pull_request:46953", + "left": "pull_request:43451", + "right": "pull_request:46891", "accept": false, - "reason": "FSDP2 migration vs default-dtype removal in norms; unrelated." + "reason": "Model addition vs CI workflow change; not the same change." }, { - "left": "pull_request:43451", + "left": "pull_request:45144", "right": "pull_request:46707", "accept": false, - "reason": "Model addition vs FSDP2 distributed-module migration; unrelated." + "reason": "Model addition vs FSDP2 integration work; unrelated." }, { - "left": "pull_request:46707", - "right": "pull_request:46891", + "left": "pull_request:46935", + "right": "pull_request:46961", "accept": false, - "reason": "FSDP2 migration vs CI-only PEFT integration changes; different scope." + "reason": "Trackio callback fix vs workflow normalization; unrelated." }, { - "left": "pull_request:46707", - "right": "pull_request:46836", + "left": "pull_request:43451", + "right": "pull_request:46658", "accept": false, - "reason": "FSDP2 migration vs OCR model addition; unrelated." + "reason": "Different new-model additions (Molmo2 vs Step-3.7-Flash), not duplicates." }, { - "left": "pull_request:46707", + "left": "pull_request:46920", "right": "pull_request:46957", "accept": false, - "reason": "FSDP2 migration vs generation/FP8 optimization; different code paths." + "reason": "mistral-common API adaptation for Voxtral Realtime vs generation decode optimization; unrelated." }, { - "left": "pull_request:46836", - "right": "pull_request:46957", + "left": "pull_request:46957", + "right": "pull_request:46958", "accept": false, - "reason": "Unlimited OCR model addition vs FP8 decode optimization; unrelated." + "reason": "One is generation optimization, the other is a CohereASR loss bugfix; different code paths." }, { "left": "pull_request:46836", - "right": "pull_request:46891", + "right": "pull_request:46920", "accept": false, - "reason": "OCR model addition vs CI workflow tweak; not duplicates." + "reason": "Unlimited OCR model addition vs Voxtral Realtime API adaptation; unrelated." }, { "left": "pull_request:46891", - "right": "pull_request:46957", + "right": "pull_request:46958", "accept": false, - "reason": "CI workflow edits vs generation/FP8 optimization; unrelated." + "reason": "CI/workflow work vs CohereASR training-loss fix; different changes." }, { - "left": "pull_request:46891", - "right": "pull_request:46953", + "left": "pull_request:46957", + "right": "pull_request:46965", "accept": false, - "reason": "CI workflow change vs norm dtype cleanup; different concerns." + "reason": "Generation optimization vs DeepSeek V4 cache semantics; not the same bug." }, { - "left": "pull_request:46903", - "right": "pull_request:46958", + "left": "pull_request:46953", + "right": "pull_request:46968", "accept": false, - "reason": "Both touch CohereASR, but #46903 is a broader multi-model loss-shift fix and not the same concrete change as the CohereASR-only bug fix." + "reason": "Model dtype cleanup vs docs autolink fixes; unrelated." }, { - "left": "pull_request:46658", - "right": "pull_request:46903", + "left": "pull_request:46707", + "right": "pull_request:46957", "accept": false, - "reason": "Step-3.7-Flash model work vs CohereASR/PPFormulaNet loss-shift fix; unrelated." + "reason": "FSDP2 migration vs generation optimization; different infrastructure areas." }, { - "left": "pull_request:46903", - "right": "pull_request:46964", + "left": "pull_request:46196", + "right": "pull_request:46957", "accept": false, - "reason": "Training-loss shift fixes vs quantized PEFT validation; different bugs." + "reason": "Wav2vec2 SDPA support vs generation optimization; unrelated." }, { "left": "pull_request:46912", "right": "pull_request:46957", "accept": false, - "reason": "RoPE tests vs FP8 decode optimization; test coverage is not the same change." + "reason": "RoPE testing improvements vs generation optimization; different intent and code." }, { - "left": "pull_request:43451", - "right": "pull_request:46912", + "left": "pull_request:46957", + "right": "pull_request:46968", "accept": false, - "reason": "Model addition vs RoPE test improvements; unrelated." + "reason": "Generation optimization vs docs autolink cleanup; unrelated." }, { - "left": "pull_request:46891", - "right": "pull_request:46912", + "left": "pull_request:46850", + "right": "pull_request:46957", "accept": false, - "reason": "CI workflow vs RoPE tests; different changes." + "reason": "Gemma sliding-window attention masking bugfix vs generation optimization; different underlying issue." }, { - "left": "pull_request:46836", + "left": "pull_request:43451", "right": "pull_request:46912", "accept": false, - "reason": "OCR model addition vs RoPE test changes; unrelated." + "reason": "Molmo2 model addition vs RoPE tests; unrelated." }, { - "left": "pull_request:46707", - "right": "pull_request:46912", + "left": "pull_request:46850", + "right": "pull_request:46891", "accept": false, - "reason": "FSDP2 migration vs RoPE tests; unrelated." + "reason": "Gemma masking bugfix vs CI workflow changes; unrelated." }, { - "left": "pull_request:46935", + "left": "pull_request:46891", + "right": "pull_request:46968", + "accept": false, + "reason": "CI workflow change vs docs autolink cleanup; unrelated." + }, + { + "left": "pull_request:43451", "right": "pull_request:46957", "accept": false, - "reason": "Trackio callback fix vs FP8 generation optimization; unrelated." + "reason": "Molmo2 model addition vs generation optimization; unrelated." }, { - "left": "pull_request:46891", - "right": "pull_request:46935", + "left": "pull_request:46196", + "right": "pull_request:46836", "accept": false, - "reason": "CI workflow change vs Trackio callback fix; different code paths." + "reason": "Wav2vec2 SDPA dispatch vs Unlimited OCR model addition; unrelated." }, { "left": "pull_request:43451", - "right": "pull_request:46935", + "right": "pull_request:46836", "accept": false, - "reason": "Model addition vs Trackio callback fix; unrelated." + "reason": "Different model additions (Molmo2 vs Unlimited OCR); not the same underlying change." }, { - "left": "pull_request:46935", - "right": "pull_request:46961", + "left": "pull_request:46658", + "right": "pull_request:46836", "accept": false, - "reason": "Trackio callback bug vs Serge branch normalization workflow; unrelated." + "reason": "Step-3.7-Flash model build vs Unlimited OCR model addition; unrelated." }, { - "left": "pull_request:46935", - "right": "pull_request:46953", + "left": "pull_request:43451", + "right": "pull_request:46196", "accept": false, - "reason": "Trackio callback fix vs dtype cleanup in fused norms; unrelated." + "reason": "Molmo2 model addition vs wav2vec2 SDPA dispatch; unrelated." }, { "left": "pull_request:46953", "right": "pull_request:46957", "accept": false, - "reason": "RMSNorm dtype cleanup vs FP8 decode optimization; unrelated." + "reason": "Model dtype cleanup vs generation optimization; unrelated." }, { - "left": "pull_request:46836", - "right": "pull_request:46953", + "left": "pull_request:46891", + "right": "pull_request:46957", "accept": false, - "reason": "OCR model addition vs dtype cleanup; unrelated." + "reason": "CI integration work vs generation optimization; unrelated." }, { - "left": "pull_request:46954", - "right": "pull_request:46961", + "left": "pull_request:46707", + "right": "pull_request:46836", + "accept": false, + "reason": "FSDP2 migration vs Unlimited OCR model addition; unrelated." + }, + { + "left": "pull_request:46961", + "right": "pull_request:46965", + "accept": false, + "reason": "Workflow normalization vs DeepSeek V4 cache support; unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted duplicate pair is the CohereASR training-loss double-shift fix (46895/46958), which is supported by the shared files and shared issue target. The other verdicts appropriately treat surface similarity as insufficient evidence for duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 36, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43451-24", + "nodes": [ + "pull_request:43451", + "pull_request:46196", + "pull_request:46707", + "pull_request:46836", + "pull_request:46850", + "pull_request:46891", + "pull_request:46912", + "pull_request:46953", + "pull_request:46957", + "pull_request:46961", + "pull_request:46965", + "pull_request:46968" + ], + "soft_pairs": [ + "pull_request:46707|pull_request:46850", + "pull_request:46891|pull_request:46912", + "pull_request:46836|pull_request:46912", + "pull_request:43451|pull_request:46968", + "pull_request:46707|pull_request:46912", + "pull_request:46707|pull_request:46968", + "pull_request:43451|pull_request:46965", + "pull_request:46196|pull_request:46953", + "pull_request:46836|pull_request:46965", + "pull_request:46196|pull_request:46891", + "pull_request:46957|pull_request:46961", + "pull_request:46891|pull_request:46961", + "pull_request:43451|pull_request:46961", + "pull_request:46953|pull_request:46961" + ], + "prepared_review_unit_hash": "f0094e2801060f78c9e1babd3c06b844f41dddb046a7abbe209951a25661612f", + "component_budget": { + "node_count": 24, + "item_count": 24, + "soft_pair_count": 86, + "serialized_chars": 90472, + "estimated_input_tokens": 22618, + "estimated_eval_tokens": 45492 + }, + "budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 14, + "serialized_chars": 33353, + "estimated_input_tokens": 8339, + "estimated_eval_tokens": 16934 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: model additions, unrelated bug fixes, docs, CI/workflow changes, and performance/infrastructure work. I found no true duplicate pair among the soft edges, so no single canonical PR stands out.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs do not share the same underlying bug or change; they span unrelated features and fixes across models, docs, CI, and distributed/runtime code.", + "best_issue_reason": null, + "best_pr_reason": "No global best PR for deduplication: none of these PRs is a safe representative for the cluster because they are not semantically overlapping enough to merge or substitute for one another.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46707", + "right": "pull_request:46850", + "accept": false, + "reason": "FSDP2 module/migration work vs Gemma 3/4 sliding-window attention masking; different code paths and different problems." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46912", "accept": false, - "reason": "Pytest v9 compatibility fix vs workflow normalization for Serge branches; unrelated." + "reason": "CI workflow for PEFT integration vs RoPE testing improvements; both are infra/test related but not the same change." }, { - "left": "pull_request:46891", - "right": "pull_request:46954", + "left": "pull_request:46836", + "right": "pull_request:46912", "accept": false, - "reason": "CI workflow tweak vs Pytest v9 fix; different changes." + "reason": "Unlimited OCR model addition vs RoPE test coverage; unrelated model and test work." }, { "left": "pull_request:43451", - "right": "pull_request:46954", + "right": "pull_request:46968", "accept": false, - "reason": "Model addition vs Pytest v9 compatibility fix; unrelated." + "reason": "Molmo2 model addition vs docs autolink cleanup; unrelated feature vs documentation fix." }, { - "left": "pull_request:46954", - "right": "pull_request:46957", + "left": "pull_request:46707", + "right": "pull_request:46912", "accept": false, - "reason": "Pytest compatibility fix vs FP8 generation optimization; unrelated." + "reason": "Native FSDP2 migration vs RoPE test expansion; no shared concrete bug or code path." }, { - "left": "pull_request:46953", - "right": "pull_request:46954", + "left": "pull_request:46707", + "right": "pull_request:46968", "accept": false, - "reason": "Different fixes: dtype removal in norms vs Pytest v9 compatibility." + "reason": "FSDP2 runtime/distributed changes vs docs autolink cleanup; entirely different scope." }, { - "left": "pull_request:46958", - "right": "pull_request:46961", + "left": "pull_request:43451", + "right": "pull_request:46965", "accept": false, - "reason": "CohereASR loss fix vs Serge workflow normalization; unrelated." + "reason": "Molmo2 model add vs DeepSeek V4 use_cache fix; different models and issues." }, { - "left": "pull_request:46957", - "right": "pull_request:46958", + "left": "pull_request:46196", + "right": "pull_request:46953", "accept": false, - "reason": "FP8 generation optimization vs CohereASR loss-shift bug; unrelated." + "reason": "wav2vec2 SDPA dispatch support vs removing default dtype in FusedRMSNormGated modules; unrelated model/runtime changes." }, { - "left": "pull_request:46891", - "right": "pull_request:46958", + "left": "pull_request:46836", + "right": "pull_request:46965", "accept": false, - "reason": "CI workflow tweak vs CohereASR loss fix; unrelated." + "reason": "Unlimited OCR model addition vs DeepSeek V4 cache handling; different models and distinct fixes." }, { - "left": "pull_request:43451", - "right": "pull_request:46958", + "left": "pull_request:46196", + "right": "pull_request:46891", "accept": false, - "reason": "Model addition vs CohereASR training-loss fix; unrelated." + "reason": "wav2vec2 SDPA dispatch/model changes vs PEFT CI workflow; not the same underlying change." }, { - "left": "pull_request:46836", - "right": "pull_request:46964", + "left": "pull_request:46957", + "right": "pull_request:46961", "accept": false, - "reason": "Unlimited OCR model addition vs PEFT quantized-training validation; unrelated." + "reason": "FP8 generation kernel optimization vs Serge branch-normalization workflow; unrelated changes." }, { - "left": "pull_request:46658", - "right": "pull_request:46964", + "left": "pull_request:46891", + "right": "pull_request:46961", "accept": false, - "reason": "Step-3.7-Flash model work vs PEFT quantized-training validation; unrelated." + "reason": "PEFT integration CI work vs Serge fix normalization workflow; both touch CI but solve different problems." }, { - "left": "pull_request:46707", - "right": "pull_request:46964", + "left": "pull_request:43451", + "right": "pull_request:46961", "accept": false, - "reason": "FSDP2 migration vs PEFT quantized-training validation; unrelated." + "reason": "Molmo2 model addition vs automated patch-normalization workflow; unrelated." }, { - "left": "pull_request:46895", - "right": "pull_request:46958", - "accept": true, - "reason": "Same CohereASR training-loss double-shift bug, same file set, and the fixes are alternative implementations of the same code-path correction." + "left": "pull_request:46953", + "right": "pull_request:46961", + "accept": false, + "reason": "FusedRMSNormGated dtype cleanup vs Serge normalization workflow; no shared bug or mergeable code path." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative. It correctly treats most pairs as unrelated broad-similarity false positives and only elevates the CohereASR pair that is explicitly tied to the same issue target and same file area. The soft-edge rejects are appropriately cautious; no pair is overstated as a duplicate beyond the evidence provided." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -22446,52 +22574,52 @@ "pull_request:43488|pull_request:43532", "pull_request:43532|pull_request:46747" ], - "prepared_review_unit_hash": "b5d4b8fc877b1251284d0831b31d44832644b9a7f3a3901735b4a9bd99297fe2", + "prepared_review_unit_hash": "3c346fea6026ace24b0ef8113944e7f5873423c4c5ba8c735d60388c57c89522", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 6347, - "estimated_input_tokens": 1587, - "estimated_eval_tokens": 3430 + "serialized_chars": 6350, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 6347, - "estimated_input_tokens": 1587, - "estimated_eval_tokens": 3430 + "serialized_chars": 6350, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three items are unrelated \"don't merge\" test PRs: one tweaks setup/versioning and a Vit test, one changes CI/comment-workflow plumbing across several files, and one flips a security-gate workflow trigger. There is no same underlying bug or change cluster here.", - "confidence": 0.96, + "summary": "These three PRs are all synthetic '[don't merge]' changes, but they do not appear to fix the same underlying problem. They span unrelated areas: packaging/versioning, PR-comment/workflow automation, and a security-gate workflow change.", + "confidence": 0.36, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the three PRs target different code paths and workflows, so none is a duplicate of the others.", + "canonical_pr_reason": "PR #43532 is the broadest and most representative of the CI/testing automation changes in this set, touching multiple workflow and test-related files.", "best_issue_reason": null, - "best_pr_reason": "If one representative PR must be chosen, #43532 is the broadest and most consequential change set, but it is still not a duplicate match for the others.", + "best_pr_reason": "PR #43532 is the strongest standalone representative because it has the widest scope and the most context across workflow and testing utilities.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:46747", "accept": false, - "reason": "Different changes: #43488 is about setup/dependency/version and a test file, while #46747 only changes a security-gate workflow trigger. Same broad repo, not the same fix." + "reason": "One is a packaging/versioning and test-dependency change; the other only flips a workflow trigger for the security gate. Different code paths and different fixes." }, { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "They share only a test filename, but the actual edits are unrelated: #43488 adjusts setup/versioning and test scaffolding, while #43532 rewires multiple workflows and helper utilities." + "reason": "They overlap only on a Vit test file, but the actual changes are unrelated: one is setup/dependency/versioning, the other is workflow automation plus test helpers." }, { "left": "pull_request:43532", "right": "pull_request:46747", "accept": false, - "reason": "Both are workflow-related, but they modify different CI paths and problems: #43532 changes comment/self-scheduled/slack handling, while #46747 changes PR security-gate triggering. Not mergeable as one fix." + "reason": "Both touch workflows, but they address different CI paths and different behavior: self-comment/slack/scheduled automation versus the PR security gate." } ] }, @@ -22521,12 +22649,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "d3887a07c42c8a7631df309b136ffe2e59a50c4a48aaf3ad6d63f3c46910cb39", + "prepared_review_unit_hash": "bccfaddb20abd42e853295edafbe3b6ee6eee267825d3a6a3a6dd01a240f47d7", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6083, + "serialized_chars": 6082, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -22534,7 +22662,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6083, + "serialized_chars": 6082, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -22543,18 +22671,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate pipeline-addition PRs: one for Promptable Concept Segmentation and one for Promptable Visual Segmentation. They share framework plumbing and docs patterns, but they target different tasks/models and should not be merged as duplicates.", + "summary": "These are two separate PRs adding different segmentation pipelines: concept segmentation vs visual segmentation. They share scaffolding and docs, but target different tasks and code paths, so they should not be deduplicated.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: the two pull requests implement different segmentation pipelines with different task scopes and model integrations.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "Neither PR is a duplicate of the other. If forced to pick a representative, 43613 is broader in model coverage, but it is not a better canonical for 43612 because the underlying feature is different.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Reject: they add different concrete pipelines (concept segmentation vs visual segmentation) with different filenames, docs, and model hookups. Shared infrastructure changes are not enough to treat them as the same change." + "reason": "Different features and implementations: Promptable Concept Segmentation vs Promptable Visual Segmentation. Shared docs/pipeline boilerplate is not enough to treat them as the same fix." } ] }, @@ -22584,40 +22712,40 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "ff33b0dc7ff009c73c0b4f162f07c8604dc539a04fc188c869ada04868d584cf", + "prepared_review_unit_hash": "d6196c36fa4e3b9207533e0a7b2683ebaedb7be0e7942e6afb2d2ce1143f68b5", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4292, - "estimated_input_tokens": 1073, - "estimated_eval_tokens": 2402 + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4292, - "estimated_input_tokens": 1073, - "estimated_eval_tokens": 2402 + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs in `trainer.py` that both improve custom-trainer extensibility, but they address different code paths: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering. They are not duplicates or mergeable into one PR as-is.", - "confidence": 0.98, + "summary": "Both PRs touch `src/transformers/trainer.py`, but they address different trainer extensibility points: gradient-accumulation loss scaling vs. overriding signature-column filtering. They are related in theme, not the same underlying change or bug.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: these are separate trainer enhancements affecting different behavior (`training_step` loss scaling vs column-filter signature inspection).", + "canonical_pr_reason": "PR 43651 is the cleaner representative of the cluster because it introduces a concrete behavior fix in `training_step` with a narrowly scoped API hook; PR 43663 is a separate refactor/extension for column filtering.", "best_issue_reason": null, - "best_pr_reason": "PR 43651 is the clearest self-contained behavior change, but neither PR should be treated as a duplicate of the other.", + "best_pr_reason": "PR 43651 is the best single representative: it has the more direct bug-fix/change semantics and a self-contained code-path impact, while 43663 is an independent trainer customization change.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Both touch `src/transformers/trainer.py`, but they fix different concerns: 43651 changes loss scaling during gradient accumulation, while 43663 extracts signature-column logic for filtering. Different code paths, no shared concrete bug, and they would not plausibly be merged into one PR." + "reason": "Different code paths and problems: 43651 changes loss scaling during gradient accumulation; 43663 changes how trainer signature columns are derived for filtering. Shared file/subsystem alone is not enough to treat them as duplicates." } ] }, @@ -22647,12 +22775,12 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "8f02cbf8c7fa0b660658e75954a17ec354e43c1c06a052ded8af553de92f1778", + "prepared_review_unit_hash": "1e9b7f4b82b2b4abaaff88e99d9d07b319b0aa2de447ad3873500d10bc056154", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3664, + "serialized_chars": 3663, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -22660,7 +22788,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3664, + "serialized_chars": 3663, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -22669,18 +22797,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two pull requests touch the same CLI file and both mention TypeAdapter, but they fix different runtime/type-annotation issues. The tighter, more bug-specific PR is the pydantic-optional guard; the other is a broader annotation cleanup.", - "confidence": 0.82, + "summary": "Two PRs touch the same CLI file and both mention TypeAdapter, but they fix different NameError causes and are not safe duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 is the more precise representative: it directly guards `TypeAdapter` behind `TYPE_CHECKING` and quotes the annotation, matching the reported no-pydantic NameError scenario more closely than the broader typing cleanup in 43656.", + "canonical_pr_reason": "PR 43842 is the more specific, concrete fix: it addresses the runtime NameError when pydantic is absent by guarding the TypeAdapter annotation/import. PR 43656 is a broader typing cleanup around annotations in the same file.", "best_issue_reason": null, - "best_pr_reason": "PR 43842 is the best single artifact in the cluster because it targets a concrete runtime failure path with a minimal, focused fix.", + "best_pr_reason": "PR 43842 is the best representative of the cluster because it targets a clear runtime failure mode in the CLI. PR 43656 is related but solves a different annotation-related problem.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Same file and similar `TypeAdapter` wording, but the fixes are not the same underlying change: 43656 is a broader future-annotations/type-hint cleanup, while 43842 specifically handles `TypeAdapter` when pydantic is absent. Too different to treat as duplicates." + "reason": "Same file and same symbol, but different bugs: one is a general annotation/type-hint NameError fix, the other specifically prevents importing TypeAdapter when pydantic is missing. They are not the same underlying change." } ] }, @@ -22872,394 +23000,640 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "4d4ba6923136013b3581d2d6fbac4db8b8fb29cecb5f0759a662f98fbf48f4ac", + "prepared_review_unit_hash": "67531753cef017df60c4b6a2937f1adbc6cb8813a8a79374976431a7dccb3bce", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84897, - "estimated_input_tokens": 21225, - "estimated_eval_tokens": 42706 + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 }, "budget": { "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75094, - "estimated_input_tokens": 18774, - "estimated_eval_tokens": 37804 + "serialized_chars": 75098, + "estimated_input_tokens": 18775, + "estimated_eval_tokens": 37806 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly separate model-specific refactors under the same output-tracing initiative. The only clear duplicate groups are the GPT-Neo pair and the GPT-J/CodeGen trio; most other pairs share only the same broad tracking issue or decorator pattern.", - "confidence": 0.87, - "canonical_issue_reason": null, - "canonical_pr_reason": "44066 is the strongest representative of the duplicate family: it is explicitly tied to #43979, covers the broad GPT-J/CodeGen refactor, and has the most central scope among the GPT-J-related PRs.", - "best_issue_reason": null, - "best_pr_reason": "44066 is the best global representative of this cluster because it is broad, explicitly issue-linked, and sits at the center of the only real duplicate subgroup in the set.", + "summary": "This cluster is one umbrella output-tracing refactor campaign tied to issue #43979, with a few true duplicates where PRs touch the same concrete model code path (especially GPT-J / CodeGen and GPT-Neo variants), but most pairs are just same-pattern refactors across different models and should stay separate.", + "confidence": 0.95, + "canonical_issue_reason": "#43979 is the shared umbrella tracking issue referenced by the refactor PRs and best represents the common output-tracing work.", + "canonical_pr_reason": "PR #44722 is the strongest canonical representative: it covers the same GPT-J/CodeGen output-tracing refactor as the overlapping PRs and appears to be the most complete/latest version of that concrete change.", + "best_issue_reason": "#43979 is the only issue-level umbrella in the cluster and is the natural anchor for the duplicated refactor PRs.", + "best_pr_reason": "#44722 is the best overall PR representative for the duplicated GPT-J/CodeGen refactor slice of this cluster.", "soft_edge_verdicts": [ { - "left": "pull_request:43996", - "right": "pull_request:44085", + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models and code paths; same issue style only." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different model families; not the same concrete bug/change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both target GPT-J output tracing on the same code path; 44066 is just a broader variant that also includes CodeGen." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "Both are output-tracing refactors, but for different vision backbones and different implementations." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet vs GPT-J/CodeGen; unrelated concrete changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "Different models and forward paths." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa are unrelated code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen refactor vs MGP-STR refactor; different model code." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "Different model architectures and forward implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both are GPT-Neo output-tracing refactors touching the same model code path and look mergeable as one PR." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen versus GPT-Neo; same theme, not the same change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "Different models and different concrete code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet refactor and GPT-J/CodeGen refactor are unrelated changes." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are different model paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated concrete implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different model code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet vs MPT; not the same underlying bug/change." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo refactor is separate from GPT-J/CodeGen refactor." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated implementations." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are different code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT refactor is not the same concrete change as GPT-J/CodeGen." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo versus GPT-J; same refactor pattern but different model code." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different model paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated concrete implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are unrelated model changes." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are unrelated changes." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 refactor is separate from GPT-J/CodeGen." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are unrelated implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer refactor is separate from GPT-J/CodeGen." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are different model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are unrelated concrete changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are different model implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are unrelated refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are different model code paths." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are unrelated changes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are different model implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", "accept": false, - "reason": "Different models and code paths; only the standardized output-tracing theme overlaps." + "reason": "TextNet and LongT5 are unrelated concrete refactors." }, { - "left": "pull_request:43996", + "left": "pull_request:44018", + "right": "pull_request:44161", + "accept": false, + "reason": "GPT-Neo and LongT5 are different code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are unrelated model changes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are different code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are unrelated concrete changes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are different model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "SqueezeBert and GPT-Neo are unrelated code paths." + }, + { + "left": "pull_request:44010", "right": "pull_request:44044", "accept": false, - "reason": "Different architectures (CVT/FNet vs DeBERTa); same refactor pattern is not enough." + "reason": "SqueezeBert and DeBERTa v2 are different concrete changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa v2 and TextNet are unrelated refactors." }, { "left": "pull_request:44066", - "right": "pull_request:44085", + "right": "pull_request:44722", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same code path; they look like overlapping implementations of the same change." + "reason": "Same GPT-J/CodeGen output-tracing refactor on the same concrete files; clearly duplicate territory." }, { - "left": "pull_request:44007", - "right": "pull_request:44072", + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same model path and could be merged into one PR." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", "accept": false, - "reason": "RegNet/ResNet-family refactor versus EfficientNet; different model-specific changes." + "reason": "GPT-J and SpeechT5 are unrelated changes." }, { - "left": "pull_request:44072", - "right": "pull_request:44722", + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "FNet/CVT versus ResNet/RegNet/RT-DETR; not the same bug or change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "FNet/CVT and SpeechT5 are different code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa v2 and SpeechT5 are unrelated model changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "FNet/CVT and SegFormer are different implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", "accept": false, - "reason": "EfficientNet versus GPT-J/CodeGen; unrelated concrete changes." + "reason": "FNet/CVT and LongT5 are unrelated refactors." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen versus EfficientNet; same initiative, but not the same underlying bug/change." + "reason": "GPT-J/CodeGen and EfficientNet are different model code paths." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are unrelated model implementations." + "reason": "MobileNetV2 and DeBERTa v2 are unrelated." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen refactor versus MGP-STR refactor; different code paths." + "reason": "GPT-J/CodeGen and MGP-STR are different model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen versus MPT; too broad a similarity to treat as duplicate." + "reason": "GPT-J/CodeGen and MPT are unrelated concrete changes." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both target GPT-Neo output tracing on the same files and appear to be the same refactor in different PRs." + "reason": "Same GPT-Neo output-tracing refactor; same code path and mergeable as one PR." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are separate model implementations." + "reason": "GPT-J/CodeGen versus GPT-Neo; same pattern, different model code." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo versus GPT-J/CodeGen; same theme, different model code." + "reason": "GPT-Neo and GPT-J/CodeGen are distinct model paths." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet versus GPT-J/CodeGen; unrelated model changes." + "reason": "ResNet refactor is unrelated to GPT-J/CodeGen." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are unrelated concrete changes." + "reason": "MGP-STR and GPT-J/CodeGen are different concrete changes." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet versus GPT-J; no same code path." + "reason": "ResNet and GPT-J are unrelated implementations." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are different model families." + "reason": "GPT-Neo and ResNet are different model code paths." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet versus MPT; only the refactor pattern overlaps." + "reason": "ResNet and MPT are unrelated changes." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo versus GPT-J/CodeGen; different model-specific implementations." + "reason": "GPT-Neo is separate from GPT-J/CodeGen." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo versus Flaubert; unrelated models." + "reason": "GPT-Neo and Flaubert are different concrete refactors." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo versus Flaubert; not the same underlying change." + "reason": "GPT-Neo and Flaubert/XLM are unrelated." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT versus GPT-J/CodeGen; unrelated code paths." + "reason": "MPT and GPT-J/CodeGen are different model paths." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo versus GPT-J; same broad topic only." + "reason": "GPT-Neo and GPT-J are distinct refactors." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus MGP-STR; different models and behavior." + "reason": "GPT-Neo and MGP-STR are unrelated changes." }, { "left": "pull_request:44068", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus MGP-STR; not a duplicate." + "reason": "GPT-Neo and MGP-STR are different code paths." }, { - "left": "pull_request:44059", + "left": "pull_request:44071", "right": "pull_request:44086", "accept": false, - "reason": "GPT-2 versus MGP-STR; different model-specific refactors." + "reason": "MPT and MGP-STR are unrelated model implementations." }, { - "left": "pull_request:44098", - "right": "pull_request:44101", + "left": "pull_request:44085", + "right": "pull_request:44086", "accept": false, - "reason": "ViLT versus Flaubert/XLM; unrelated implementations." + "reason": "GPT-J and MGP-STR are different changes." }, { "left": "pull_request:44059", - "right": "pull_request:44722", - "accept": false, - "reason": "GPT-2 versus GPT-J/CodeGen; same pattern, different concrete change." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44071", + "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus MPT; not the same bug or patch." + "reason": "GPT-2 and MGP-STR are unrelated." }, { - "left": "pull_request:44056", + "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "MPNet versus Flaubert/XLM; different code paths." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44017", - "accept": false, - "reason": "RegNet versus Segformer; only shared output-tracing refactor language." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44056", - "accept": false, - "reason": "RegNet versus MPNet; different model-specific changes." + "reason": "ViLT and Flaubert/XLM are different code paths." }, { - "left": "pull_request:44071", - "right": "pull_request:44076", + "left": "pull_request:44059", + "right": "pull_request:44722", "accept": false, - "reason": "MPT versus ImageGPT; unrelated model implementations." + "reason": "GPT-2 refactor is unrelated to GPT-J/CodeGen." }, { - "left": "pull_request:43997", - "right": "pull_request:44161", + "left": "pull_request:44018", + "right": "pull_request:44071", "accept": false, - "reason": "RegNet versus LongT5; no shared code path." + "reason": "GPT-Neo and MPT are different model code paths." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "Segformer versus LongT5; same initiative, different models." + "reason": "SegFormer and LongT5 are unrelated refactors." }, { "left": "pull_request:44017", "right": "pull_request:44722", "accept": false, - "reason": "Segformer versus GPT-J/CodeGen; unrelated changes." + "reason": "SegFormer and GPT-J/CodeGen are unrelated." }, { "left": "pull_request:44017", "right": "pull_request:44073", "accept": false, - "reason": "Segformer versus VisualBert; different model code." - }, - { - "left": "pull_request:44073", - "right": "pull_request:44101", - "accept": false, - "reason": "VisualBert versus Flaubert/XLM; not the same underlying fix." + "reason": "SegFormer and VisualBert are different model implementations." }, { - "left": "pull_request:44073", - "right": "pull_request:44161", + "left": "pull_request:44068", + "right": "pull_request:44071", "accept": false, - "reason": "VisualBert versus LongT5; different implementations." + "reason": "GPT-Neo and MPT are unrelated concrete changes." }, { - "left": "pull_request:44059", - "right": "pull_request:44071", + "left": "pull_request:44071", + "right": "pull_request:44076", "accept": false, - "reason": "GPT-2 versus MPT; different model-specific output handling." + "reason": "MPT and ImageGPT are different model paths." }, { - "left": "pull_request:44056", - "right": "pull_request:44161", + "left": "pull_request:44074", + "right": "pull_request:44101", "accept": false, - "reason": "MPNet versus LongT5; unrelated concrete changes." + "reason": "TextNet and Flaubert/XLM are unrelated changes." }, { "left": "pull_request:43997", "right": "pull_request:44018", "accept": false, - "reason": "RegNet versus GPT-Neo; not duplicates." + "reason": "RegNet and GPT-Neo are different model code paths." }, { - "left": "pull_request:44071", - "right": "pull_request:44074", + "left": "pull_request:44056", + "right": "pull_request:44161", "accept": false, - "reason": "MPT versus TextNet; different model code." + "reason": "MPNet and LongT5 are unrelated refactors." }, { "left": "pull_request:43997", - "right": "pull_request:44129", + "right": "pull_request:44161", "accept": false, - "reason": "RegNet versus SpeechT5; same tracking theme only." + "reason": "RegNet and LongT5 are different concrete changes." }, { - "left": "pull_request:43996", + "left": "pull_request:44074", "right": "pull_request:44161", "accept": false, - "reason": "CVT/FNet versus LongT5; unrelated refactors." + "reason": "TextNet and LongT5 are unrelated." }, { - "left": "pull_request:43996", - "right": "pull_request:44017", + "left": "pull_request:44018", + "right": "pull_request:44161", + "accept": false, + "reason": "GPT-Neo and LongT5 are different model paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", "accept": false, - "reason": "CVT/FNet versus Segformer; different models." + "reason": "RegNet and TextNet are unrelated model changes." }, { "left": "pull_request:44010", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBert versus MPT; no shared code path." + "reason": "SqueezeBert and MPT are different code paths." }, { - "left": "pull_request:44010", - "right": "pull_request:44073", + "left": "pull_request:44018", + "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert versus VisualBert; unrelated model-specific changes." + "reason": "GPT-Neo and SpeechT5 are unrelated refactors." }, { "left": "pull_request:44010", - "right": "pull_request:44018", + "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert versus GPT-Neo; different implementations." + "reason": "SqueezeBert and VisualBert are different model implementations." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", + "left": "pull_request:44010", + "right": "pull_request:44018", "accept": false, - "reason": "TextNet versus Flaubert/XLM; not the same change." + "reason": "SqueezeBert and GPT-Neo are unrelated changes." }, { "left": "pull_request:44010", "right": "pull_request:44044", "accept": false, - "reason": "SqueezeBert versus DeBERTa; different code paths." + "reason": "SqueezeBert and DeBERTa v2 are different concrete code paths." }, { "left": "pull_request:44044", "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa versus TextNet; unrelated model refactors." - }, - { - "left": "pull_request:44044", - "right": "pull_request:44129", - "accept": false, - "reason": "DeBERTa versus SpeechT5; same pattern only." - }, - { - "left": "pull_request:44044", - "right": "pull_request:44154", - "accept": false, - "reason": "DeBERTa versus VITS; unrelated concrete implementations." - }, - { - "left": "pull_request:44066", - "right": "pull_request:44722", - "accept": true, - "reason": "Both are GPT-J/CodeGen output-tracing refactors on the same code path and could plausibly be merged into one PR." - }, - { - "left": "pull_request:44085", - "right": "pull_request:44722", - "accept": true, - "reason": "Both appear to be GPT-J output-tracing refactors with overlapping file scope; same underlying change." - }, - { - "left": "pull_request:44085", - "right": "pull_request:44129", - "accept": false, - "reason": "GPT-J versus SpeechT5; shared issue tag is not enough." - }, - { - "left": "pull_request:43996", - "right": "pull_request:44007", - "accept": false, - "reason": "CVT/FNet versus RegNet/ResNet-family; different model-specific refactors." - }, - { - "left": "pull_request:43996", - "right": "pull_request:44129", - "accept": false, - "reason": "CVT/FNet versus SpeechT5; unrelated concrete changes." + "reason": "DeBERTa v2 and TextNet are unrelated." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa versus VITS; different code paths and model families." + "reason": "DeBERTa v2 and VITS are different model code paths." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays within the packet: it correctly identifies the GPT-Neo pair and the GPT-J/CodeGen subgroup as the only clear duplicate-like clusters, and it treats the rest conservatively as same-initiative but different model-specific refactors. The accepted soft-edge pairs are mergeable as single bugfix PRs, and the rejected pairs avoid overstating similarity." + "feedback": "Grounded in the packet: #43979 is the clear umbrella issue, and #44722 is a reasonable canonical PR for the GPT-J/CodeGen slice because it is the latest/most complete overlapping change. The soft-edge calls are conservative overall: the only true-duplicate verdicts are for the same model-path refactors (GPT-J/CodeGen and GPT-Neo), while the rest are appropriately separated by model/code path." }, "evaluator_used": true, "retried": false, @@ -23297,70 +23671,70 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "419929bace3d0a62292aa3fc03fdff0487271047e412b43bbcc8a4e71f72192d", + "prepared_review_unit_hash": "4cadede3443298b9f5cd151b2903f29c2ace5152c0423413cee974d44059f093", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84897, - "estimated_input_tokens": 21225, - "estimated_eval_tokens": 42706 + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 6, - "serialized_chars": 17124, - "estimated_input_tokens": 4281, - "estimated_eval_tokens": 8818 + "serialized_chars": 17125, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all part of a broad wave of model-specific refactors to standardized output tracing/capture, but they target different architectures and code paths. They are related by technique, not duplicates.", - "confidence": 0.93, + "summary": "This cluster is a family of model-specific refactors to the standardized output-capturing / tracing APIs (capture_outputs, can_return_tuple, _can_record_outputs). They are thematically similar but mostly independent changes across different model implementations, not duplicate fixes for one underlying bug.", + "confidence": 0.74, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44073 is the clearest representative of the shared refactor pattern: it explicitly updates output tracing with both capture_outputs and can_return_tuple in a typical model implementation.", + "canonical_pr_reason": "PR 44129 is the best anchor for the cluster: it has an explicit issue target (43979) and a concrete SpeechT5 output-tracing refactor, making it the most clearly grounded representative change.", "best_issue_reason": null, - "best_pr_reason": "PR 44073 is the best overall exemplar of the cluster\u2019s theme because its title and diff most directly reflect the standardized output-capture migration, while still being a concrete single-model change.", + "best_pr_reason": "PR 44129 is the strongest representative of the shared migration theme because it is explicitly tied to an issue and updates a concrete code path rather than just following the broader refactor pattern.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Both are standardized output-capture refactors, but they touch different model implementations (FocalNet vs ImageGPT) and different concrete code paths." + "reason": "Both are output-tracing migrations, but they touch different model code paths (FocalNet vs ImageGPT) with model-specific hooks and decorator wiring. Same general refactor theme, not the same underlying bug/change." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "VisualBert and TextNet are separate model-specific refactors; the shared decorator migration is not enough to make them the same change." + "reason": "VisualBert and TextNet are separate model implementations. The diffs show similar decorator migration work, but no shared concrete fix or mergeable single PR scope." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet and VisualBert changes are parallel refactors in different architectures, not one mergeable duplicate fix." + "reason": "These are parallel output-capture refactors in different models (MPNet vs VisualBert). They are related by infrastructure, not by a single bug or code-path correction." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "These update different model files and model internals; they are similar refactors but not the same underlying bug or PR." + "reason": "MPNet and TextNet are distinct model-specific refactors; the overlap is only the standardized tracing pattern, which is too broad to treat as duplicate work." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert changes are model-specific output-tracing migrations, not the same concrete code-path problem." + "reason": "RegNet and SqueezeBert changes are both migration-style edits, but they affect different model internals and do not appear to fix the same concrete defect." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 are unrelated model implementations; they share a refactor theme but not a duplicate code change." + "reason": "SqueezeBert and SpeechT5 are unrelated model code paths. Shared use of output-capturing utilities is insufficient to merge them as one change." } ] }, @@ -23440,274 +23814,274 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "2fa800ff2efa0c6ca4c7d531517fe3563a6ccf6e9bec2dbf4f2cf891891b6bde", + "prepared_review_unit_hash": "88de7bc6d7ab46b44b67dbd9179e353d1bb79c54e375c965fc62f99279ee84e8", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40170, - "estimated_input_tokens": 10043, - "estimated_eval_tokens": 20342 + "serialized_chars": 40183, + "estimated_input_tokens": 10046, + "estimated_eval_tokens": 20348 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40170, - "estimated_input_tokens": 10043, - "estimated_eval_tokens": 20342 + "serialized_chars": 40183, + "estimated_input_tokens": 10046, + "estimated_eval_tokens": 20348 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "A cluster of independent, model-specific output-tracing refactors. They share the same general pattern, but they do not look like duplicates of the same concrete bug or change.", - "confidence": 0.93, + "summary": "This is a loose cluster of independent, model-specific \"refactor output tracing\" pull requests. They share a common refactoring theme, but they target different architectures and code paths, so they should not be merged as duplicates.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44003 is the broadest representative of the shared output-tracing refactor pattern, touching two related model files and introducing the new capture/output-recording plumbing explicitly.", + "canonical_pr_reason": "PR 44003 is the most representative of the cluster: it makes the broadest output-tracing change, spans two related model files, and touches the shared output-capturing plumbing rather than only a single wrapper.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 is the strongest cluster representative: it has the most substantial refactor and shows the new output-capturing pattern clearly, but it is still not a duplicate anchor for the other model-specific PRs.", + "best_pr_reason": "PR 44003 is the strongest global representative because it is the most infrastructure-level output-tracing refactor in the set and is broader than the single-model adaptations.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model files and code paths (timm_backbone vs upernet); same refactor theme, not the same change." + "reason": "Same refactor theme, but `timm_backbone` and `upernet` are unrelated model paths with different output logic." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different architectures (timm_backbone vs vision_text_dual_encoder); not a single concrete bug or mergeable change." + "reason": "Both mention output tracing, but these are different model families and different concrete changes." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different model implementations (timm_backbone vs univnet); only the output-tracing pattern matches." + "reason": "Unrelated architectures (`timm_backbone` vs `univnet`); not the same underlying bug or change." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Different families and code paths (timm_backbone vs mamba/falcon_mamba); not the same underlying bug/change." + "reason": "Different code paths and mechanisms; one is a backbone refactor, the other is mamba output-capture plumbing." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different model code paths (timm_backbone vs rwkv); similar refactor, not a duplicate." + "reason": "Different model families and different forward/output semantics; too broad to be duplicates." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "Different models (mobilenet_v1 vs univnet); same refactor style, but no shared concrete fix." + "reason": "Both are output-tracing refactors, but they affect different model implementations and are not mergeable as one PR." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different architectures (mobilenet_v1 vs upernet); not the same code-path problem." + "reason": "Different model code paths (`mobilenet_v1` vs `upernet`), so this is only thematic similarity." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "Different model families (mobilenet_v1 vs depth_anything/prompt_depth_anything); same pattern only." + "reason": "Separate model families with distinct tracing changes; not the same concrete fix." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different model code paths (mobilenet_v1 vs vision_encoder_decoder); not mergeable as one PR." + "reason": "`mobilenet_v1` and `vision_encoder_decoder` change different forward APIs and output handling." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "Different model code paths (mobilenet_v1 vs speech_encoder_decoder); only broadly similar refactoring." + "reason": "Different encoder-decoder vs CNN model code paths; only the refactor theme overlaps." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Different architectures (vision_text_dual_encoder vs univnet); same output-tracing theme, not same change." + "reason": "`vision_text_dual_encoder` and `univnet` are unrelated; the output-tracing changes are model-specific." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different model files and behaviors (vision_text_dual_encoder vs upernet); not a duplicate." + "reason": "Different model subsystems and different output shapes/return conventions." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Different code paths (vision_text_dual_encoder vs mamba/falcon_mamba); too broad a similarity to merge." + "reason": "These touch unrelated architectures and different tracing implementations, not one shared bug." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different model families (vision_text_dual_encoder vs depth_anything/prompt_depth_anything); same pattern only." + "reason": "Different model families and separate output-tracing refactors; not the same concrete change." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "Different model implementations (univnet vs upernet); no shared concrete bug/change." + "reason": "Different models (`univnet` vs `upernet`) and different forward signatures; too broad to merge." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different architectures and code paths (univnet vs mamba/falcon_mamba); not the same PR-worthy fix." + "reason": "`univnet` and `mamba` fix different output-tracing paths in unrelated code." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Different model files (univnet vs depth_anything/prompt_depth_anything); same refactor theme, not duplicate work." + "reason": "Separate model families; same refactor motif is not enough for duplicate triage." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "Different models (upernet vs rwkv); only similar output-tracing cleanup, not a shared bug." + "reason": "`upernet` and `rwkv` are different architectures with unrelated output handling." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different code paths (upernet vs mamba/falcon_mamba); not one concrete change." + "reason": "Different model stacks and different implementation details; not the same underlying change." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "Different architectures (mamba/falcon_mamba vs superpoint); same tracing refactor pattern, not the same fix." + "reason": "`mamba`/`falcon_mamba` output-capture plumbing is unrelated to `superpoint`'s refactor." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Different model code paths (mamba/falcon_mamba vs codegen); too different to count as duplicates." + "reason": "`codegen` changes cache/attention forward typing; this is a different code-path change than mamba tracing." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Different model families (timm_backbone vs codegen); same broad refactor theme only." + "reason": "Different model families and different refactor targets; not the same bug or feature." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "Different model implementations (univnet vs codegen); not the same underlying bug/change." + "reason": "`univnet` and `codegen` are unrelated, and the diffs alter different forward semantics." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "Different architectures (upernet vs codegen); similar output-tracing cleanup, but not duplicate." + "reason": "`upernet` and `codegen` have unrelated output-tracing/caching changes." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "Different code paths (vision_text_dual_encoder vs codegen); too broad to merge as one PR." + "reason": "Different architectures and different concrete output-path changes." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Different model families (depth_anything/prompt_depth_anything vs vision_encoder_decoder); same pattern only." + "reason": "Both are tracing refactors, but they target different model families (`depth_anything` vs `vision_encoder_decoder`)." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Different model families (depth_anything/prompt_depth_anything vs speech_encoder_decoder); not one concrete bug." + "reason": "Different model families and different forward/output conventions." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Different model code paths (depth_anything/prompt_depth_anything vs superpoint); only structurally similar." + "reason": "`depth_anything` and `superpoint` are separate model implementations with distinct changes." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Different model families (depth_anything/prompt_depth_anything vs rwkv); same refactor pattern, not duplicate." + "reason": "Different architectures and different tracing mechanisms; not mergeable as one PR." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Different model code paths (depth_anything/prompt_depth_anything vs DPR); not the same underlying change." + "reason": "`depth_anything` and `dpr` modify unrelated model outputs and signatures." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "Different model families (vision_encoder_decoder vs rwkv); similar tracing refactor only." + "reason": "`vision_encoder_decoder` and `rwkv` are unrelated subsystems with different output handling." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "Different code paths (vision_encoder_decoder vs speech_encoder_decoder); not the same concrete fix." + "reason": "Different encoder-decoder wrappers for different modalities; not the same change." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "Different models (vision_encoder_decoder vs superpoint); same theme, different change." + "reason": "`vision_encoder_decoder` vs `superpoint` are different code paths and output formats." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different model code paths (vision_encoder_decoder vs DPR); not a single duplicate PR." + "reason": "These touch different model families and different wrapper behavior." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Different models (speech_encoder_decoder vs superpoint); only similar refactor style." + "reason": "Speech encoder-decoder and SuperPoint have unrelated forward/output refactors." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Different architectures (speech_encoder_decoder vs rwkv); not the same bug or change." + "reason": "Different models and different output-tracing mechanisms." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Different code paths (speech_encoder_decoder vs DPR); too different to merge as one PR." + "reason": "`speech_encoder_decoder` and `dpr` are unrelated code paths." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "Different model implementations (superpoint vs rwkv); same output-tracing theme only." + "reason": "`superpoint` and `rwkv` are different architectures; the changes are only superficially similar." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "Different model families (superpoint vs DPR); not the same concrete fix." + "reason": "Different model implementations and different output semantics." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "Different model code paths (rwkv vs DPR); similar refactor pattern, not duplicate work." + "reason": "`rwkv` output tracing and `dpr` encoder wrapping are unrelated concrete changes." } ] }, @@ -23737,7 +24111,7 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "389c98f57f59e33993d8f0cde2e079ca1692cd971cb56e2575429c6486502ec0", + "prepared_review_unit_hash": "da3d2c27c3d69ed13c3b89eb712dc1eecd11e84e7d910c7f01770742d5d27ed9", "component_budget": { "node_count": 2, "item_count": 2, @@ -23759,18 +24133,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are both model-addition PRs, but they target different models and different code paths, so they are not duplicates.", + "summary": "These are two separate model-implementation PRs for different architectures (PenguinVL vs Voxtral TTS), not duplicates. They share some common Transformers plumbing and docs, but they fix/add different concrete code paths.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: PR #44662 adds PenguinVL, while PR #45401 adds Voxtral TTS; they are unrelated model implementations.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "No single best PR for this cluster because the PRs address different model families rather than the same underlying change.", + "best_pr_reason": "PR 45401 is the cleaner representative of the cluster only in the sense that it is more focused and explicitly tied to issue 45027, but it is not a duplicate of 44662.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Reject: these are different model implementations (PenguinVL vs Voxtral TTS) with distinct architecture, files, and generated APIs; they do not appear to fix or add the same underlying code-path change." + "reason": "Different model additions with different purposes and code paths: PenguinVL implementation vs Voxtral TTS support. Shared auto/docs files are generic registration work, not evidence of the same change." } ] }, @@ -23800,12 +24174,12 @@ "soft_pairs": [ "pull_request:44965|pull_request:46845" ], - "prepared_review_unit_hash": "6c13d5646f2d92a82ec739c9eb7d691efb317ff1c4a185ba999a69262ec12556", + "prepared_review_unit_hash": "9fe18532c9b7e6bbf94eb214d0b71edb108805a69147dc5b802423e1e2d8bd2b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4219, + "serialized_chars": 4218, "estimated_input_tokens": 1055, "estimated_eval_tokens": 2366 }, @@ -23813,7 +24187,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4219, + "serialized_chars": 4218, "estimated_input_tokens": 1055, "estimated_eval_tokens": 2366 }, @@ -23822,18 +24196,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated pull requests: one is an experimental workflow edit to change PR handling, and the other is a CI branch/test routing change plus an unrelated temporary code tweak. They do not appear to fix the same underlying problem or implement the same change.", - "confidence": 0.95, + "summary": "The two pull requests are not duplicates: one is a draft/CI workflow experiment to convert PRs to draft, while the other changes CI branch routing and includes an unrelated placeholder code edit. They do not appear to fix the same concrete problem.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No clear canonical PR: #44965 is a malformed workflow experiment, while #46845 is a separate CI branch-selection tweak plus an unrelated placeholder code change; they are not the same change.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "#46845 is the stronger standalone candidate because it makes a concrete CI routing change, whereas #44965 looks like an incomplete workflow experiment with no clear finished fix.", "soft_edge_verdicts": [ { "left": "pull_request:44965", "right": "pull_request:46845", "accept": false, - "reason": "Different code paths and intents: PR 44965 rewrites the PR documentation workflow into a draft-conversion workflow, while PR 46845 changes PR CI branch routing and includes an unrelated modeling_utils temp edit. They are not the same bug/change and would not be merged into one PR." + "reason": "Reject: the PRs target different behaviors. #44965 aims to convert PRs to draft via a workflow rewrite, while #46845 changes CI branch handling and includes an unrelated code placeholder edit; they are not the same underlying bug or mergeable fix." } ] }, @@ -23863,7 +24237,7 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "57b7f969f1b3e6d211e581ae7ddf84baa1decbe71aec776221f74dc13333f7dc", + "prepared_review_unit_hash": "43cfed12d1b60f998b66bd65d5de06bc081c4afdd54a20f85ff3b4352c28aebc", "component_budget": { "node_count": 2, "item_count": 2, @@ -23885,18 +24259,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are related but not duplicates: PR 44979 introduces a generic Module Fusion API, while PR 45363 uses KernelConfig to implement a specific n-to-1 kernel fusion workflow across integrations/modeling code. They target different layers and can plausibly land independently.", - "confidence": 0.96, + "summary": "The PRs are related but not duplicates: one introduces a generic module-fusion API, while the other integrates KernelConfig-based n-to-1 kernel fusion into model loading and adds an example.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR 44979 is the broader foundational change: it introduces the generic module_fusion API and tests. PR 45363 is a narrower, draft integration on top of that idea, so 44979 is the better representative of the underlying work.", "best_issue_reason": null, - "best_pr_reason": "PR 44979 is the cleaner canonical representative for the cluster because it defines the underlying fusion API and test coverage, whereas PR 45363 is a higher-level feature that appears to build on that foundation.", + "best_pr_reason": "PR 44979 is the strongest representative in this cluster because it is the core, non-draft API change and is broader than the specific KernelConfig fusion integration in 45363.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Different scope and code paths: 44979 adds a new generic module fusion API, while 45363 wires KernelConfig into existing model/integration code for a specific n-to-1 kernel fusion feature. They are related, but not the same underlying change and would not be merged as one PR." + "reason": "Related feature area, but not the same concrete change. 44979 adds a new generic module-fusion API; 45363 wires KernelConfig into loading for a specific n-to-1 fusion path and includes an example. They could be merged separately." } ] }, @@ -23926,7 +24300,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "37250eb3ebaca16fc79e1194887b631e447d191d39bd9ce86664eaa8249dc5ef", + "prepared_review_unit_hash": "69a94bdf6f2dd83ae97ac7b128e9525d88c66c4d587c8de87b9c0033f5d3a574", "component_budget": { "node_count": 2, "item_count": 2, @@ -23948,18 +24322,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests address the same Mistral regex patch bug in the same file and target the same issue; 45317 is the more complete version because it adds tests as well as the code fix.", - "confidence": 0.98, + "summary": "Both PRs patch the same Mistral regex pre-tokenizer AttributeError in the same file and on the same lines; 45317 is the more complete version because it adds tests as well as the code fix.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the better canonical representative: it fixes the same AttributeError as 45086, targets the same issue, and also includes a test update, making it the more complete change.", + "canonical_pr_reason": "45317 is the better canonical PR: it fixes the same bug as 45086 but also includes test coverage, making it the more complete representative of the change.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best representative of the cluster because it contains the same concrete fix as 45086 plus validation coverage, so it is the strongest single PR to keep.", + "best_pr_reason": "45317 is the strongest PR in the cluster because it addresses the concrete AttributeError and adds validation tests, while 45086 is a smaller precursor of the same fix.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Accept: both PRs modify the same Mistral pre-tokenizer code path to fix the same AttributeError under the same issue target, so they are duplicates of the same underlying change." + "reason": "Same underlying code-path bug in _patch_mistral_regex, same target issue, and the diffs overlap directly; the later PR plausibly subsumes the earlier one." } ] }, @@ -23989,12 +24363,12 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "19379c36ad33fe2beb66d317058f3ab9b0f8b6b682ae5d0976619706c4954b14", + "prepared_review_unit_hash": "0153660c376a5927927682305f06a1ec4241b6c9b761ee3c01fadc2310693163", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6065, + "serialized_chars": 6066, "estimated_input_tokens": 1517, "estimated_eval_tokens": 3290 }, @@ -24002,7 +24376,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6065, + "serialized_chars": 6066, "estimated_input_tokens": 1517, "estimated_eval_tokens": 3290 }, @@ -24011,18 +24385,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related NemotronH refactor efforts, but they are not duplicates: one PR focuses on inheritance/conversion mapping for NemotronH to GraniteMoeHybrid-style handling, while the other expands NemotronH into separate dense/sparse model families with docs and auto-mapping changes.", - "confidence": 0.92, + "summary": "These PRs overlap on NemotronH internals, but they are not the same change: one is a refactor toward GraniteMoeHybrid-style inheritance/conversion mapping, while the other is a broader model split introducing NemotronHDense/NemotronHSparse with docs and tests. Treat as related, not duplicates.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45569 is the broader and more complete NemotronH redesign, introducing the dense/sparse split and updating docs and auto-mappings. PR #45115 looks like a narrower preparatory refactor around inheritance/conversion mapping.", + "canonical_pr_reason": "No true canonical duplicate here; if one must be chosen as the broader representative, PR #45569 is the more complete user-facing model change, but it is not the same underlying fix as #45115.", "best_issue_reason": null, - "best_pr_reason": "PR #45569 is the best representative for the cluster because it covers the larger architectural change and the resulting model/API surface changes, whereas #45115 is only a partial implementation of the same area.", + "best_pr_reason": "PR #45569 is the strongest standalone PR in the cluster because it includes the architecture split, model registrations, docs, and tests. PR #45115 is a narrower refactor/conversion mapping change.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "They touch overlapping NemotronH files, but the concrete changes differ: #45115 is a conversion/inheritance refactor toward GraniteMoeHybrid compatibility, while #45569 splits NemotronH into dense/sparse variants with broader docs and auto-model updates. Not the same underlying change and not safely mergeable as one PR." + "reason": "Shared NemotronH files are not enough: #45115 is a focused refactor/conversion mapping/inheritance change, while #45569 is a broader NemotronH rework that adds dense/sparse variants and documentation/tests. They do not look mergeable into one duplicate PR." } ] }, @@ -24052,7 +24426,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "531c2515aff629995823f69983cc3a5ad2a95b8a77feeb1d90b1a0134ead234c", + "prepared_review_unit_hash": "bbe087218a903ef8449c8967779824ff1adc877c61632828d0787c5018b06db1", "component_budget": { "node_count": 2, "item_count": 2, @@ -24074,18 +24448,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two closely related feature PRs, but not true duplicates: one adds heterogeneous configuration support, while the other extends that work into heterogeneous model/modeling support.", - "confidence": 0.84, + "summary": "Two related but distinct open PRs: one adds full heterogeneous model support including modeling/runtime changes, while the other is a narrower config-only subset. They overlap heavily in docs/config files but are not the same concrete change.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45332 is the broader, more complete implementation and appears to subsume the configuration-only groundwork in 45333.", + "canonical_pr_reason": "PR 45332 is the better canonical representative because it is the broader umbrella change and includes the config work plus the downstream modeling/runtime pieces. PR 45333 is a narrower config-only slice.", "best_issue_reason": null, - "best_pr_reason": "PR 45332 is the best representative of the cluster because it covers the full heterogeneous model feature rather than only the config layer.", + "best_pr_reason": "PR 45332 is the best single PR for the cluster since it captures the full heterogeneous model feature and subsumes the shared configuration work.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "They share scaffolding and filenames, but the changes are not the same concrete fix/change: 45333 is config-only, while 45332 adds the larger modeling layer as well. Related feature work, not duplicate PRs." + "reason": "Related feature work, but not the same underlying change: 45332 spans heterogeneous modeling/runtime support, while 45333 focuses only on per-layer configuration. They would not plausibly be merged as one duplicate PR without losing scope distinctions." } ] }, @@ -24115,7 +24489,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "70e5a50a07c86780db3b1a62e5d5bccd652264c9759146de09ac06ecbf22bd94", + "prepared_review_unit_hash": "4f6620af032f973cb13c41d51ff29b87e2230985bbd8622d340405e59f4730df", "component_budget": { "node_count": 2, "item_count": 2, @@ -24137,18 +24511,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are typing-focused, but they address different scopes: #45415 is a broad repo-wide type-checking sweep, while #45425 adds specific typing protocols/aliases and related modeling_utils support. They overlap in a few files, but they are not the same concrete change set and would not naturally merge as one PR.", - "confidence": 0.92, + "summary": "Both pull requests are typing-related, but they address different kinds of work: one is a broad type-checking sweep across many transformer modules, while the other adds specific typing protocols/aliases and related fixes. They are similar in theme, but not the same underlying change.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "#45415 is the broader umbrella PR for enabling type checking across many `src/transformers/*` files, so it is the better representative of the cluster.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "#45415 is the best representative PR because it covers the larger end-to-end typing sweep rather than a narrower supporting typing refactor.", + "best_pr_reason": "PR 45425 is the more focused and concrete typing change set; PR 45415 is a larger sweep of type-checking adjustments and unrelated file-by-file cleanups, so it is less representative of a single duplicate target.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Shared typing-related files alone are not enough: #45415 is a broad type-checking pass, while #45425 adds specific typing helpers/protocols and modeling_utils-related support. They are adjacent but not the same underlying change." + "reason": "Shared files and a typing theme are not enough here: 45415 is a broad mypy/type-checking pass, while 45425 introduces specific typing aliases/protocols and targeted fixes. They do not look like the same code-path bug or a single mergeable PR." } ] }, @@ -24178,12 +24552,12 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "c1c58282da316cddf3d02fcc387341adfcfb832f2c5507dac7fb089d61c99e40", + "prepared_review_unit_hash": "e4a3aa5d077dbee574cdb56d9e83214e86ef7898aa3131afdef3112b68db9304", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4508, + "serialized_chars": 4507, "estimated_input_tokens": 1127, "estimated_eval_tokens": 2510 }, @@ -24191,7 +24565,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4508, + "serialized_chars": 4507, "estimated_input_tokens": 1127, "estimated_eval_tokens": 2510 }, @@ -24200,18 +24574,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related only at a high level through expert parallelism, but they fix different loading bugs and touch different code paths. I would not merge them as duplicates.", - "confidence": 0.91, + "summary": "Both PRs are about expert-parallel model loading, but they fix different failure modes: one removes `distributed_config` from serialization and adds `has_ep`, while the other prevents parameters from being overwritten during distributed/FSDP2 loading. They are related, not duplicates.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45662 is the more concrete fix: it addresses expert parameters being overwritten during rank-0 broadcast in the FSDP2 path and includes targeted loading/runtime changes. PR #45548 is about Accelerate launch with DeepSpeed ZeRO-3 and config serialization, which is a different failure mode.", + "canonical_pr_reason": "PR #45662 is the more concrete/loading-path-specific fix, with changes in the distributed tensor and module loading flow to prevent rank-0 broadcast from overwriting experts.", "best_issue_reason": null, - "best_pr_reason": "PR #45662 is the stronger standalone candidate because it targets a specific, reproducible EP loading corruption bug with code changes across the affected path, whereas #45548 is narrower and appears to solve a separate launch/serialization issue.", + "best_pr_reason": "PR #45662 best represents the cluster because it targets a specific, reproducible bug in EP + FSDP2 loading rather than a broader configuration/serialization adjustment.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Both mention EP, but the underlying bugs differ: #45548 is about Accelerate/ZeRO-3 loading and config serialization, while #45662 fixes rank-0 broadcast overwriting experts in the FSDP2 loading path. They are not the same concrete change." + "reason": "They share the EP/loading theme, but the bugs differ: #45548 fixes config serialization and an EP detection helper, while #45662 fixes parameter clobbering during distributed shard loading. Not the same concrete change." } ] }, @@ -24241,7 +24615,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "1a8ea357b62f22608379d9796e5e8afb95d29159185caf64f7e8ead7699053c4", + "prepared_review_unit_hash": "c318cbf72ebd464069a6e844acbc7f768a1830d544eeae979b525ffab97761a0", "component_budget": { "node_count": 2, "item_count": 2, @@ -24263,18 +24637,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same xdist-related captured_info/log collision problem and overlap on the core testing_utils fix. PR 45645 looks like the more complete version because it includes the core fix plus CI artifact log handling, so it is the better canonical PR.", - "confidence": 0.94, + "summary": "Both PRs address the same xdist-related captured_info collision problem in testing_utils; PR 45645 appears to be the broader end-to-end version, adding the CI-side aggregation needed to consume multiple worker-specific logs.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "Pick PR 45645 as canonical: it contains the same underlying xdist-safe captured_info fix as 45639 and also covers the CI/workflow changes needed to preserve and aggregate the debug logs.", + "canonical_pr_reason": "PR 45645 is the better canonical representative because it includes the underlying xdist-safe logging fix plus the CI workflow update to collect multiple captured_info files, making it the more complete solution.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the best representative because it appears to subsume 45639 rather than just partially overlap with it.", + "best_pr_reason": "PR 45645 is the strongest overall PR candidate: it fixes the same captured_info collision bug and also preserves CI debug-log visibility when logs are split across workers.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Same underlying xdist collision bug for captured_info/debug logs; 45645 is a broader superset of the same fix and could plausibly be merged as one PR." + "reason": "Both PRs fix the same concrete xdist collision for captured_info/debug logs in testing_utils; 45645 is a superset that also updates CI log collection, so they can plausibly be merged into one change." } ] }, @@ -24304,40 +24678,40 @@ "soft_pairs": [ "pull_request:45720|pull_request:46259" ], - "prepared_review_unit_hash": "894b46cdd145cfb457e38555730ae056ffefafb571a11c39a9fff3e17bbd9c41", + "prepared_review_unit_hash": "a2d04d6554b3eaeefd6495a179783e45bfd20ca31d58a605fc32a6d44955047e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7234, - "estimated_input_tokens": 1809, - "estimated_eval_tokens": 3874 + "serialized_chars": 7232, + "estimated_input_tokens": 1808, + "estimated_eval_tokens": 3872 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7234, - "estimated_input_tokens": 1809, - "estimated_eval_tokens": 3874 + "serialized_chars": 7232, + "estimated_input_tokens": 1808, + "estimated_eval_tokens": 3872 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are both linter/modeling-rule changes, but they implement different checks on different code paths, so they are not duplicates.", - "confidence": 0.98, + "summary": "Two linter-related PRs, but they address different changes: TRF003 adjusts forward signatures/tuple return handling, while TRF018 changes model weight initialization across many architectures.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR 45720 is the cleaner representative if one must be chosen: it is narrower and targets a single concrete forward-API refactor across model files.", "best_issue_reason": null, - "best_pr_reason": "PR 45720 is the more focused representative of the cluster: it makes a specific can_return_tuple/return_dict change across model forwards, whereas 46259 is a separate _init_weights linter rule. It is not a duplicate of 46259, but it is the cleaner canonical-style PR among the two.", + "best_pr_reason": "PR 45720 is the better standalone representative because it is more focused and self-contained than the broader initialization-rule PR.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "Reject: TRF003 adds can_return_tuple/return_dict handling in model forward methods, while TRF018 changes _init_weights initialization logic. They affect different code paths and are not the same underlying change." + "reason": "Different concrete changes and different linter rules: TRF003 adds can_return_tuple/remove return_dict handling, while TRF018 rewrites initialization logic. Shared model filenames are incidental, not the same underlying fix." } ] }, @@ -24433,40 +24807,40 @@ "soft_pairs": [ "pull_request:45932|pull_request:46100" ], - "prepared_review_unit_hash": "e5676e95d6fa073f4a76bae15d1f79e48506a68ef67ba5ddf90c6ec2aeb868c2", + "prepared_review_unit_hash": "b2184499c222b3a0ab3f5ecdf09e302fbcfd9f99d93f25744497d95dfa605967", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5161, - "estimated_input_tokens": 1291, - "estimated_eval_tokens": 2838 + "serialized_chars": 5160, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5161, - "estimated_input_tokens": 1291, - "estimated_eval_tokens": 2838 + "serialized_chars": 5160, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: one is a broad experimental/model+CI change set, the other is a test migration focused on altclip/vit kernels and DDP test scaffolding. They do not appear to fix the same concrete bug or implement the same change.", - "confidence": 0.94, + "summary": "These two pull requests are not duplicates: one is a CI/telemetry and model-mapping change, while the other is a test migration adding kernel tests and related fixtures. They do not appear to fix the same concrete code-path problem or belong in one merged PR.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46100 is the better canonical PR for the cluster because it is the more self-contained and coherent change set; PR #45932 is explicitly a 'DO NOT MERGE' experimental Grafana/CI and model-edit bundle with unrelated edits.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #46100 is the best representative PR here: it centers on a single test-migration effort, whereas PR #45932 mixes CI workflow changes with multiple model files and looks more like a testing/instrumentation scratch PR.", + "best_pr_reason": "No single PR is a good canonical representative because the PRs are unrelated in scope and purpose: PR 45932 is about CI/secrets and model fixes, while PR 46100 is about test migration and new kernel test coverage.", "soft_edge_verdicts": [ { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "Different underlying changes: #45932 is a broad experimental PR touching CI, model code, and mappings; #46100 is a test migration adding kernel tests and DDP config. They do not look mergeable into one PR or like the same bug fix." + "reason": "Different underlying changes and code paths; one is CI/model maintenance, the other is test migration and new kernel tests. They should not be merged as duplicates." } ] }, @@ -24496,7 +24870,7 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "459535b8fe21e3b2d9fb2f8f5eefcc21954b6418576c073c31d0b7382a0ce141", + "prepared_review_unit_hash": "e1c9a3274e67a1fb1e4848b1f941058c3408f4b703c867175cb98c207f0ace0b", "component_budget": { "node_count": 2, "item_count": 2, @@ -24518,18 +24892,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are related through shared distributed-code refactors, but they are not the same underlying change. PR #46126 is a focused EP/Trainer integration patch, while PR #46269 is a broad distributed branch base spanning FSDP, integrations, model loading, CI, and more.", - "confidence": 0.95, + "summary": "Two open PRs overlap on distributed-model code, but they are not the same change: one is a focused EP + Trainer integration PR, while the other is a broad distributed branch base spanning FSDP, CI, model loading, and multiple integration areas.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46126 is the more specific, self-contained change centered on EP + Trainer integration, making it the better canonical representative if one must be chosen.", + "canonical_pr_reason": "PR #46126 is the better canonical representative because it is a focused, self-contained feature PR centered on EP + Trainer integration; #46269 is too broad and appears to bundle many unrelated distributed changes.", "best_issue_reason": null, - "best_pr_reason": "PR #46126 is the best representative of the cluster because it targets one concrete feature path, whereas PR #46269 is a wide umbrella/base branch and not a single fix.", + "best_pr_reason": "PR #46126 is the strongest standalone PR in this cluster: it has a clear single purpose and coherent scope, whereas #46269 is a large umbrella branch-base change that is not suitable as a duplicate target.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Reject: the overlap is mostly due to shared distributed-module refactoring. #46126 focuses on EP routing/Trainer integration, while #46269 is a broad base branch touching many unrelated distributed subsystems, so they are not the same concrete change." + "reason": "They share some distributed filenames, but the scopes differ materially. #46126 is a specific EP + Trainer integration change; #46269 is a broad distributed branch base touching FSDP, CI, model loading, and other subsystems. They do not look like the same concrete bug or change and would not plausibly merge into one PR." } ] }, @@ -24595,208 +24969,208 @@ "pull_request:46504|pull_request:46563", "pull_request:46504|pull_request:46716" ], - "prepared_review_unit_hash": "2ca7be192013f0e2dbde3914838b435b8cd31dc4ed852f7e9c932f291447a15c", + "prepared_review_unit_hash": "88a1d1cd228e06bc1cf1be3ecc4fb0fb82711d89edd1fde944ab8f07377c4f0f", "component_budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 29, - "serialized_chars": 25933, - "estimated_input_tokens": 6484, - "estimated_eval_tokens": 13224 + "serialized_chars": 25932, + "estimated_input_tokens": 6483, + "estimated_eval_tokens": 13222 }, "budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 29, - "serialized_chars": 25933, - "estimated_input_tokens": 6484, - "estimated_eval_tokens": 13224 + "serialized_chars": 25932, + "estimated_input_tokens": 6483, + "estimated_eval_tokens": 13222 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are mostly unrelated workflow maintenance changes. The only tight subcluster is several edits to pr-ci-security-gate.yml, but each addresses a different behavior (permissions, approval flow, reusable-workflow invocation, or security checks), so none should be merged as duplicates.", - "confidence": 0.96, + "summary": "Mostly unrelated workflow chore PRs. The only tight cluster is pr-ci-security-gate.yml, but those PRs change different security-gate behaviors, so none are true duplicates.", + "confidence": 0.88, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "PR 46554 is the most substantive representative because it changes the actual security-gate logic, while the others are smaller permission/checkout/secrets plumbing tweaks in different workflows.", "soft_edge_verdicts": [ { "left": "pull_request:46138", "right": "pull_request:46554", "accept": false, - "reason": "Different workflow changes: stdin handling for self-comment CI vs adding security checks to pr-ci-security-gate.yml." + "reason": "Different workflows and different problems: stdin handling for self-comment CI vs security-gate logic." }, { "left": "pull_request:46138", "right": "pull_request:46558", "accept": false, - "reason": "Different workflow changes in different code paths; not the same bug or fix." + "reason": "Unrelated workflow changes; one updates comment parsing, the other adjusts PR CI gate permissions/approval flow." }, { "left": "pull_request:46138", "right": "pull_request:46563", "accept": false, - "reason": "Different workflow changes; one is self-comment CI shell input handling, the other is a security-gate reusable-workflow comment/change." + "reason": "Different code paths and purposes; not the same underlying fix." }, { "left": "pull_request:46138", "right": "pull_request:46716", "accept": false, - "reason": "Different workflow changes; unrelated implementation and intent." + "reason": "No shared bug/change beyond both being workflow edits." }, { "left": "pull_request:46138", "right": "pull_request:46504", "accept": false, - "reason": "Self-comment CI stdin handling is unrelated to build-ci-docker-images commit-message handling." + "reason": "Self-comment CI message piping vs build CI docker tag detection; unrelated." }, { "left": "pull_request:46432", "right": "pull_request:46503", "accept": false, - "reason": "SSH runner hardening/memory-monitor changes are unrelated to benchmark trigger conditions." + "reason": "SSH runner hardening is unrelated to benchmark dispatch support." }, { "left": "pull_request:46432", "right": "pull_request:46501", "accept": false, - "reason": "Different workflows and different behaviors; not the same underlying change." + "reason": "Different workflows and different concerns; no duplicate bug/change." }, { "left": "pull_request:46432", "right": "pull_request:46502", "accept": false, - "reason": "SSH runner workflow changes do not match assign-reviewers checkout-ref changes." + "reason": "SSH runner changes do not match reviewer assignment workflow changes." }, { "left": "pull_request:46432", "right": "pull_request:46554", "accept": false, - "reason": "Unrelated workflow files and fixes." + "reason": "Different workflows and different fixes; not mergeable as one change." }, { "left": "pull_request:46432", "right": "pull_request:46558", "accept": false, - "reason": "Unrelated workflow files and fixes." + "reason": "SSH runner hardening vs PR CI gate permissions/approval logic are unrelated." }, { "left": "pull_request:46501", "right": "pull_request:46503", "accept": false, - "reason": "Both touch workflows, but anti-slop permissions and benchmark dispatch triggering are different changes." + "reason": "Both are workflow chores, but they change different files and different behavior." }, { "left": "pull_request:46501", "right": "pull_request:46502", "accept": false, - "reason": "Different workflows and different behavior changes." + "reason": "Anti-slop permission scoping and reviewer assignment ref checkout are different changes." }, { "left": "pull_request:46501", "right": "pull_request:46504", "accept": false, - "reason": "Anti-slop permissions update is unrelated to build-ci-docker-images commit-message parsing." + "reason": "Anti-slop workflow vs build CI docker image trigger logic; unrelated." }, { "left": "pull_request:46501", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different fixes." + "reason": "Different workflow files and different security/automation behavior." }, { "left": "pull_request:46502", "right": "pull_request:46503", "accept": false, - "reason": "Assign-reviewers checkout ref change is unrelated to benchmark_v2 trigger logic." + "reason": "Different workflows and unrelated behavior changes." }, { "left": "pull_request:46502", "right": "pull_request:46504", "accept": false, - "reason": "Different workflow files and different fixes." + "reason": "Reviewer assignment ref checkout is unrelated to build image trigger parsing." }, { "left": "pull_request:46502", "right": "pull_request:46554", "accept": false, - "reason": "Different workflows and different behaviors." + "reason": "No shared underlying bug; reviewer workflow vs security-gate workflow." }, { "left": "pull_request:46503", "right": "pull_request:46504", "accept": false, - "reason": "Benchmark dispatch logic and build-ci-docker-images message handling are distinct changes." + "reason": "Benchmark dispatch support and build-image trigger parsing are different concerns." }, { "left": "pull_request:46503", "right": "pull_request:46554", "accept": false, - "reason": "Different workflow behavior and code paths." + "reason": "Different workflow and different code-path problem." }, { "left": "pull_request:46504", "right": "pull_request:46554", "accept": false, - "reason": "Build-ci-docker-images commit-message parsing is unrelated to pr-ci-security-gate security checks." + "reason": "Build-ci-docker-images trigger parsing vs PR security gate logic are unrelated." }, { "left": "pull_request:46504", "right": "pull_request:46558", "accept": false, - "reason": "Different workflow changes in different files." + "reason": "Different workflow files and different behavior." }, { "left": "pull_request:46554", "right": "pull_request:46558", "accept": false, - "reason": "Same file, but different security-gate changes: one adds file-change/security checks, the other adjusts permissions and approval flow." + "reason": "Same file, but different fixes: one adds diff-based security checks; the other changes permissions and approval gating." }, { "left": "pull_request:46554", "right": "pull_request:46563", "accept": false, - "reason": "Same workflow file, but one adds security checks while the other changes reusable-workflow invocation/comments; not the same fix." + "reason": "Same workflow family, but different underlying changes: local checks vs reusable workflow migration." }, { "left": "pull_request:46554", "right": "pull_request:46716", "accept": false, - "reason": "Same workflow file, but one adds checks and the other changes secrets inheritance; distinct behavior." + "reason": "Different security hardening steps: local checks vs secrets scoping." }, { "left": "pull_request:46558", "right": "pull_request:46563", "accept": false, - "reason": "Different pr-ci-security-gate behavior changes; not a single concrete fix." + "reason": "Different PR CI gate modifications with different code paths and intent." }, { "left": "pull_request:46558", "right": "pull_request:46716", "accept": false, - "reason": "Permission tightening/approval logic is different from secrets removal in the reusable workflow call." + "reason": "Permissions/approval fix is not the same as changing reusable workflow secrets handling." }, { "left": "pull_request:46563", "right": "pull_request:46716", "accept": false, - "reason": "Different changes: safety comments on the reusable workflow vs changing secrets passing to {}." + "reason": "Both touch the same workflow file, but one changes workflow source and the other removes inherited secrets; distinct fixes." }, { "left": "pull_request:46504", "right": "pull_request:46563", "accept": false, - "reason": "Different workflow files and unrelated fixes." + "reason": "Build image trigger parsing vs security-gate workflow migration are unrelated." }, { "left": "pull_request:46504", "right": "pull_request:46716", "accept": false, - "reason": "Different workflow files and unrelated fixes." + "reason": "No common underlying change beyond being workflow edits." } ] }, @@ -24889,12 +25263,12 @@ "soft_pairs": [ "pull_request:46310|pull_request:46546" ], - "prepared_review_unit_hash": "fab65f424ba1c06971f02c8a15e27386d74f986f07b78f17c3a2747aa67f5906", + "prepared_review_unit_hash": "d17fa1e3a8dc4af4c01a1475171e08a83da7a28e6662ffec301a9f7ec75bc370", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4851, + "serialized_chars": 4850, "estimated_input_tokens": 1213, "estimated_eval_tokens": 2682 }, @@ -24902,7 +25276,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4851, + "serialized_chars": 4850, "estimated_input_tokens": 1213, "estimated_eval_tokens": 2682 }, @@ -24911,18 +25285,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs fix the same sliding-window right-boundary bug (`-dist <= right_window_size`) in Moonshine streaming; 46546 is a broader, more complete version that also patches Gemma4 and adds tests.", - "confidence": 0.84, + "summary": "Both PRs fix the same sliding-window right-boundary off-by-one bug in Moonshine streaming; #46546 is the broader, more complete patch and also includes the related Gemma4 fix and tests.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46546 is the more complete representative: it includes the same Moonshine streaming boundary fix as 46310, adds test coverage, and bundles the related Gemma4 boundary correction.", + "canonical_pr_reason": "PR #46546 is the better canonical PR because it contains the Moonshine streaming boundary fix plus the matching Gemma4 update and test coverage, making it the most complete representative of this bug-fix cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 46546 best captures the underlying change set because it fixes the same concrete boundary condition in Moonshine streaming and strengthens it with tests; 46310 is just the narrower precursor.", + "best_pr_reason": "PR #46546 is the best representative overall because it addresses the same concrete boundary bug as #46310, but with broader coverage and validation.", "soft_edge_verdicts": [ { "left": "pull_request:46310", "right": "pull_request:46546", "accept": true, - "reason": "Same right-window boundary fix in the same Moonshine streaming code path; 46546 is a superset that plausibly subsumes 46310's change." + "reason": "Same underlying off-by-one in the right-window mask; #46546 is effectively the superset patch with the same Moonshine fix." } ] }, @@ -24944,87 +25318,72 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-46376-6", + "cluster_id": "cluster-46376-5", "nodes": [ "pull_request:46376", "pull_request:46417", "pull_request:46425", "pull_request:46509", - "pull_request:46850", "pull_request:46917" ], "soft_pairs": [ "pull_request:46376|pull_request:46417", "pull_request:46376|pull_request:46425", "pull_request:46425|pull_request:46509", - "pull_request:46425|pull_request:46917", - "pull_request:46425|pull_request:46850", - "pull_request:46850|pull_request:46917" + "pull_request:46425|pull_request:46917" ], - "prepared_review_unit_hash": "c2a3517ed934f560195532decf5505d592078dbdedc9ce420e541fe6d6aa6d64", + "prepared_review_unit_hash": "4ec1bb05fc23428202bfd951785ab2a3a3f8f251fc70752a38e45827ceefb853", "component_budget": { - "node_count": 6, - "item_count": 6, - "soft_pair_count": 6, - "serialized_chars": 16831, - "estimated_input_tokens": 4208, - "estimated_eval_tokens": 8672 + "node_count": 5, + "item_count": 5, + "soft_pair_count": 4, + "serialized_chars": 13785, + "estimated_input_tokens": 3447, + "estimated_eval_tokens": 7150 }, "budget": { - "node_count": 6, - "item_count": 6, - "soft_pair_count": 6, - "serialized_chars": 16831, - "estimated_input_tokens": 4208, - "estimated_eval_tokens": 8672 + "node_count": 5, + "item_count": 5, + "soft_pair_count": 4, + "serialized_chars": 13785, + "estimated_input_tokens": 3447, + "estimated_eval_tokens": 7150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are mostly unrelated: multimodal continuous batching, two separate model additions, one Gemma masking bug fix, and one cleanup/deprecation removal. None of the soft pairs look like the same underlying change.", + "summary": "This cluster is heterogeneous: one continuous-batching multimodal feature PR, three separate model-integration PRs, and one deprecation cleanup PR. They do not look like duplicate reports of the same underlying change.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46850 is the strongest canonical representative because it is a focused, concrete bug fix in a core code path (attention masking), unlike the others which are model additions or broad feature/cleanup work.", + "canonical_pr_reason": "PR 46425 is the best single representative if one must be chosen: it is a self-contained model integration with docs, auto-mapping updates, and tests, and it is not a draft.", "best_issue_reason": null, - "best_pr_reason": "PR 46850 is the best standalone PR in this set because it addresses a specific functional defect with clear code-path impact and is the most self-contained change.", + "best_pr_reason": "PR 46425 is the most complete and conventional addition in the set; the others are either draft, narrower, or unrelated cleanup.", "soft_edge_verdicts": [ { "left": "pull_request:46376", "right": "pull_request:46417", "accept": false, - "reason": "Different changes: continuous batching multimodal support vs. adding the HunYuan VL model package. They touch different code paths and are not mergeable as one fix." + "reason": "Different changes and codepaths: multimodal continuous-batching infrastructure vs adding the HunYuan VL model package, config, processor, and docs." }, { "left": "pull_request:46376", "right": "pull_request:46425", "accept": false, - "reason": "Both are model/feature work in different areas, but one is continuous batching and the other is a new RADIO model implementation. No shared concrete bug or change." + "reason": "Different features: continuous-batching multimodal support vs RADIO model integration; they are not the same bug or change." }, { "left": "pull_request:46425", "right": "pull_request:46509", "accept": false, - "reason": "Separate model additions for RADIO and Nemotron Omni. Similar scope as model support, but not the same underlying code change or bug." - }, - { - "left": "pull_request:46425", - "right": "pull_request:46917", - "accept": false, - "reason": "RADIO model support and deprecation cleanup are unrelated; they do not fix the same code-path problem." + "reason": "Both are model-related but for different architectures and implementations (RADIO vs Nemotron Omni). They would not plausibly be merged into one PR as a single fix/change." }, { "left": "pull_request:46425", - "right": "pull_request:46850", - "accept": false, - "reason": "A new RADIO model implementation and a Gemma attention-masking bug fix are unrelated changes in different subsystems." - }, - { - "left": "pull_request:46850", "right": "pull_request:46917", "accept": false, - "reason": "Gemma masking logic fix and deprecated-args cleanup are distinct changes with no common concrete defect." + "reason": "Unrelated scope: new RADIO model support versus removing deprecated training/image-processing args. Similar repository area is not enough to treat them as duplicates." } ] }, @@ -25054,12 +25413,12 @@ "soft_pairs": [ "pull_request:46574|pull_request:46580" ], - "prepared_review_unit_hash": "d58f330bddff109ac32c9c399000d2d7b45d8c0f08299b05d8948f18af2f18f4", + "prepared_review_unit_hash": "093f867ad3150fc37af1691781aebc37414ada8597f47434f874a1910c5ee219", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4410, + "serialized_chars": 4409, "estimated_input_tokens": 1103, "estimated_eval_tokens": 2462 }, @@ -25067,7 +25426,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4410, + "serialized_chars": 4409, "estimated_input_tokens": 1103, "estimated_eval_tokens": 2462 }, @@ -25076,18 +25435,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate Transformers tokenizer fixes, not duplicates: one fixes CodeLlama decode stripping of real leading spaces, the other fixes DeepSeek-Coder/DeepSeek-LLM loading via the wrong tokenizer class in auto-tokenization. They affect different code paths and would not plausibly merge into one PR.", - "confidence": 0.97, + "summary": "The two PRs are both tokenizer-related but address different bugs in different code paths: one fixes CodeLlama decode stripping, the other fixes DeepSeek checkpoint-to-tokenizer-class resolution. They are not the same underlying change.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #46574 is the more self-contained, concrete bugfix: it targets a single tokenizer decoder behavior in one model family and has a narrowly scoped code change.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:46574", "right": "pull_request:46580", "accept": false, - "reason": "Different underlying bugs and different fixes: CodeLlama decode-space stripping vs. DeepSeek checkpoint-to-tokenizer-class selection. They touch unrelated code paths and are not mergeable as one PR." + "reason": "Different models, different failure modes, and different fixes: CodeLlama decode-space stripping vs DeepSeek auto-tokenizer class misloading. They would not plausibly be merged into one PR." } ] }, @@ -25117,7 +25476,7 @@ "soft_pairs": [ "pull_request:46603|pull_request:46604" ], - "prepared_review_unit_hash": "2d7a8a4428fe22e60b4df7679f3958e9267e31d98cd156968636562bd5e9c19a", + "prepared_review_unit_hash": "c2e4cff7505a676873262bb712444840534433031ef73908b3bdac6c73ee3ed5", "component_budget": { "node_count": 2, "item_count": 2, @@ -25139,18 +25498,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open Mistral PRs overlap in files but address different changes: one relocates the existing MistralConverter into the integrations package, while the other adds tekken tokenizer detection/conversion/save support. They are related refactoring/features, not the same fix.", - "confidence": 0.91, + "summary": "Both PRs are Mistral tokenizer-related, but they target different changes: one moves the existing MistralConverter into the integrations package, while the other adds tekken tokenizer detection/conversion/save utilities. They share files because they touch the same subsystem, not because they fix the same underlying problem.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46604 is the better canonical PR because it represents the more substantive functional change for Mistral tokenization support, whereas PR 46603 is mainly an internal code move/refactor.", + "canonical_pr_reason": "No single canonical PR: 46603 is a refactor/package move, while 46604 introduces separate tekken tokenizer functionality.", "best_issue_reason": null, - "best_pr_reason": "PR 46604 is the strongest overall candidate: it adds new tokenizer detection, conversion, and save utilities, making it more representative of the cluster\u2019s user-facing Mistral work.", + "best_pr_reason": "46604 is the more substantial and self-contained change, so it is the better representative PR for this cluster.", "soft_edge_verdicts": [ { "left": "pull_request:46603", "right": "pull_request:46604", "accept": false, - "reason": "Do not merge: 46603 only moves MistralConverter into a new package, while 46604 adds tekken tokenizer support and utilities. They touch the same area but implement different concrete changes and are not the same underlying PR." + "reason": "Reject: they are adjacent Mistral tokenizer changes, but 46603 only relocates MistralConverter whereas 46604 adds tekken tokenizer detection/conversion/save utilities. Different code-path problem, not mergeable as one duplicate PR." } ] }, @@ -25211,184 +25570,184 @@ "pull_request:46908|pull_request:46950", "pull_request:46926|pull_request:46950" ], - "prepared_review_unit_hash": "a30196dd704221d7de4ef42a272fbffe3efa31ce6e33dff83ec1ac7f84be88f6", + "prepared_review_unit_hash": "94716d05c63fadf07139b5456dc1eea68ff91ae51ad953e4525fee0a8e3bb7eb", "component_budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 25, - "serialized_chars": 24931, - "estimated_input_tokens": 6233, - "estimated_eval_tokens": 12722 + "serialized_chars": 24934, + "estimated_input_tokens": 6234, + "estimated_eval_tokens": 12724 }, "budget": { "node_count": 9, "item_count": 9, "soft_pair_count": 25, - "serialized_chars": 24931, - "estimated_input_tokens": 6233, - "estimated_eval_tokens": 12722 + "serialized_chars": 24934, + "estimated_input_tokens": 6234, + "estimated_eval_tokens": 12724 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs only share a broad \u201cfix integration tests\u201d pattern; they target different models, different failure modes, and different code paths. I do not see any true duplicate PRs here, so the soft links should be treated as false positives.", - "confidence": 0.96, + "summary": "This cluster is heterogeneous: each PR targets a different model or subsystem, and the apparent overlaps are mostly superficial test-fix patterns rather than the same underlying bug or change.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46949 is the strongest representative only in the loose sense that it contains concrete model runtime fixes, but it is still unrelated to the other PRs as a duplicate.", + "canonical_pr_reason": "No true duplicate PR stands out; if a representative is needed, 46949 is the most central by similarity and impact, but the set is too mixed to form a single canonical change.", "best_issue_reason": null, - "best_pr_reason": "PR 46949 is the best standalone representative of the set because it makes actual forward-pass bug fixes in model code, whereas most others are test-only or registry tweaks.", + "best_pr_reason": "46941 is the cleanest standalone fix: it corrects a specific DAC distance-calculation bug in model code with a direct output-mismatch symptom.", "soft_edge_verdicts": [ { "left": "pull_request:46907", "right": "pull_request:46926", "accept": false, - "reason": "Different models and fixes: EdgeTam kwarg handling in model code vs MiniMax-M3-VL load setup in tests." + "reason": "Different models and different failure contexts; Edgetam kwarg cleanup is unrelated to MiniMax test load workarounds." }, { "left": "pull_request:46907", "right": "pull_request:46941", "accept": false, - "reason": "Unrelated code paths: EdgeTam processor kwarg cleanup vs DAC codebook distance formula." + "reason": "Edgetam processor-kwarg handling and DAC distance math fix are unrelated code paths." }, { "left": "pull_request:46907", "right": "pull_request:46908", "accept": false, - "reason": "Same broad theme of test failures, but different models and different fixes; one is model code, the other is test offload setup." + "reason": "Unrelated fixes: Edgetam model kwargs vs DeepSeek test offload folder for loading." }, { "left": "pull_request:46907", "right": "pull_request:46927", "accept": false, - "reason": "Different underlying bugs in different models: EdgeTam kwarg filtering vs GLM46V input_id casting." + "reason": "Different models and bugs; no shared concrete code-path problem." }, { "left": "pull_request:46907", "right": "pull_request:46928", "accept": false, - "reason": "Model code fix vs tokenizer auto-registration list change; not the same bug." + "reason": "Edgetam runtime kwarg cleanup is unrelated to BLIP-2 tokenizer registration." }, { "left": "pull_request:46908", "right": "pull_request:46926", "accept": false, - "reason": "Both add offload_folder in tests, but they are separate model load issues and not one shared bug." + "reason": "Both use offload_folder in tests, but they are separate model suites and not the same underlying bug/change." }, { "left": "pull_request:46908", "right": "pull_request:46928", "accept": false, - "reason": "Load-error test workaround vs BLIP-2 tokenizer mapping update; unrelated changes." + "reason": "DeepSeek loading workaround and BLIP-2 tokenizer mapping fix are unrelated." }, { "left": "pull_request:46908", "right": "pull_request:46951", "accept": false, - "reason": "Separate failures in different subsystems: DeepSeekV32 loading vs PEFT hotswap mismatch handling." + "reason": "Different subsystems and failure modes; test offload workaround is not PEFT hotswap loading logic." }, { "left": "pull_request:46908", "right": "pull_request:46941", "accept": false, - "reason": "Different model/test fix vs DAC numerical bug; no shared code path." + "reason": "Separate model test workaround vs DAC model arithmetic fix." }, { "left": "pull_request:46926", "right": "pull_request:46941", "accept": false, - "reason": "MiniMax-M3-VL offload test fix and DAC distance computation are unrelated." + "reason": "MiniMax test loading workaround and DAC output-mismatch bug are unrelated." }, { "left": "pull_request:46926", "right": "pull_request:46928", "accept": false, - "reason": "MiniMax load-error workaround vs BLIP-2 tokenizer registry tweak; not duplicates." + "reason": "Different model families and different fixes; no common code-path." }, { "left": "pull_request:46926", "right": "pull_request:46927", "accept": false, - "reason": "Different models and failure modes: load_error test setup vs GLM46V runtime input handling." + "reason": "MiniMax offload-folder test fix is unrelated to GLM46V input-id handling." }, { "left": "pull_request:46927", "right": "pull_request:46949", "accept": false, - "reason": "Both touch model code, but on different models and different defects; not mergeable as one duplicate fix." + "reason": "GLM46V input-id casting and audio/whisper cache-state initialization are different bugs." }, { "left": "pull_request:46927", "right": "pull_request:46951", "accept": false, - "reason": "GLM46V input dtype fix is unrelated to PEFT hotswap loading behavior." + "reason": "Unrelated model-input casting fix vs PEFT hotswap mismatch handling." }, { "left": "pull_request:46927", "right": "pull_request:46950", "accept": false, - "reason": "GLM46V forward-path dtype fix and Mamba2 OOM-related change are different bugs." + "reason": "GLM46V dtype/input handling is unrelated to Mamba2 OOM/kernel behavior." }, { "left": "pull_request:46927", "right": "pull_request:46941", "accept": false, - "reason": "Different code paths: GLM46V input_ids casting vs DAC latent decoding math." + "reason": "Different models and different concrete bugs; not mergeable as one PR." }, { "left": "pull_request:46928", "right": "pull_request:46941", "accept": false, - "reason": "Tokenizer auto mapping entry for BLIP-2 ITM is unrelated to DAC's numerical mismatch." + "reason": "Tokenizer auto-mapping change is unrelated to DAC distance computation." }, { "left": "pull_request:46928", "right": "pull_request:46951", "accept": false, - "reason": "BLIP-2 tokenizer support change and PEFT mismatched-size loading are unrelated." + "reason": "BLIP-2 tokenizer pattern addition and PEFT loading fix are unrelated." }, { "left": "pull_request:46949", "right": "pull_request:46951", "accept": false, - "reason": "AudioFlamingo3/Whisper cache handling fixes are unrelated to PEFT adapter loading behavior." + "reason": "Audio/Whisper cache-state initialization and PEFT hotswap loading are different code paths." }, { "left": "pull_request:46949", "right": "pull_request:46950", "accept": false, - "reason": "Different models and symptoms: cache state initialization vs Mamba2 OOM/combined kernel behavior." + "reason": "Different model fixes: cache-state initialization vs Mamba2 execution/oom behavior." }, { "left": "pull_request:46907", "right": "pull_request:46949", "accept": false, - "reason": "EdgeTam processor kwarg filtering and Whisper/AudioFlamingo3 cache fixes do not share the same bug." + "reason": "No shared underlying bug; Edgetam kwarg filtering is unrelated to audio/Whisper cache handling." }, { "left": "pull_request:46908", "right": "pull_request:46949", "accept": false, - "reason": "DeepSeekV32 load test workaround is unrelated to Whisper/AudioFlamingo3 runtime fixes." + "reason": "Different model test workaround vs audio/Whisper runtime fix." }, { "left": "pull_request:46950", "right": "pull_request:46951", "accept": false, - "reason": "Mamba2 OOM/kernel behavior and PEFT loading mismatch are different subsystems and failure modes." + "reason": "Mamba2 kernel/oom fix and PEFT mismatched-size loading fix are unrelated." }, { "left": "pull_request:46908", "right": "pull_request:46950", "accept": false, - "reason": "Test offload-folder workaround for DeepSeekV32 is unrelated to Mamba2's runtime/OOM fix." + "reason": "DeepSeek test load workaround and Mamba2 model/runtime change are unrelated." }, { "left": "pull_request:46926", "right": "pull_request:46950", "accept": false, - "reason": "MiniMax test-load workaround and Mamba2 kernel/OOM fix are separate issues." + "reason": "MiniMax load test workaround and Mamba2 execution fix are different problems." } ] }, @@ -25418,12 +25777,12 @@ "soft_pairs": [ "pull_request:46933|pull_request:46942" ], - "prepared_review_unit_hash": "a00a3e3148a6e866d518b3baf7dfab8032a0c0979c571ffeb168c70de5f5dc94", + "prepared_review_unit_hash": "cc88422124b3168df356e35d4c57a6b7ccf179591983300e1759016d35695390", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4372, + "serialized_chars": 4371, "estimated_input_tokens": 1093, "estimated_eval_tokens": 2442 }, @@ -25431,7 +25790,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4372, + "serialized_chars": 4371, "estimated_input_tokens": 1093, "estimated_eval_tokens": 2442 }, @@ -25440,18 +25799,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same general class of bug (incorrect dtype casting under quantization), but they fix different model families and different code paths, so they are not duplicates and should not be merged into one PR.", - "confidence": 0.97, + "summary": "Two open PRs implement the same coding pattern fix for quantized weights, but they target different model families and different code paths, so they are not duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #46933 is the better canonical representative because it has an explicit linked issue and a clearly scoped fix for Gemma4\u2019s quantized vision/audio embedders.", "best_issue_reason": null, - "best_pr_reason": "PR 46933 is the stronger representative if one must be chosen: it has an explicit issue link, and its patch clearly covers the Gemma4 vision/audio embedders with the same quantization-safe casting pattern.", + "best_pr_reason": "PR #46933 is the strongest standalone representative of the cluster: it is explicit, narrowly scoped, and clearly describes the dtype-cast bug it fixes.", "soft_edge_verdicts": [ { "left": "pull_request:46933", "right": "pull_request:46942", "accept": false, - "reason": "Same bug pattern, but different concrete implementations and models: Gemma4 vision/audio embedders vs TimesFM 2.5 ResidualBlock. Not the same underlying change and not mergeable as one PR." + "reason": "They share a general quantization dtype-cast pattern, but the concrete bugs are in different models and code paths (Gemma4 embedders vs TimesFM 2.5 ResidualBlock), with different filenames and no evidence they could be merged into one PR." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index be9e685bbc3ae8b431a1d5af9238760d0deecd21..4ef59a6550678e4b18a16a916f3b30fe94d63bdf 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260629t180016z", + "analysis_id": "hybrid-model-20260630t000016z", "archived_artifacts": { - "hybrid": "snapshots/20260629T180016Z/analysis-runs/hybrid-model-20260629t180016z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260629T180016Z/analysis-runs/hybrid-model-20260629t180016z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-06-29T18:11:41Z", + "published_at": "2026-06-30T00:09:42Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260629T180016Z", + "snapshot_id": "20260630T000016Z", "variant": "hybrid" } diff --git a/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.json b/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..76277ffafcaa1a34d5838e0ef68ebae21626218c --- /dev/null +++ b/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.json @@ -0,0 +1,2996 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260630T000016Z", + "generated_at": "2026-06-30T00:09:40Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43451-8", + "summary": "Cluster of 8 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 43451, + "issue_numbers": [], + "pr_numbers": [ + 43451, + 46658, + 46707, + 46836, + 46850, + 46891, + 46953, + 46957 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43451, + "right_pr_number": 46658, + "code_similarity": 0.243, + "size_similarity": 0.955, + "file_overlap": 0.027, + "area_overlap": 0.0, + "patch_similarity": 0.293, + "shared_filenames": [ + "src/transformers/models/auto/auto_mappings.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 385, + 392 + ], + [ + 813, + 819 + ], + [ + 933, + 939 + ], + [ + 1054, + 1060 + ], + [ + 1186, + 1192 + ] + ], + "right_ranges": [ + [ + 362, + 368 + ], + [ + 591, + 597 + ], + [ + 796, + 802 + ] + ] + } + ] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46707, + "code_similarity": 0.026, + "size_similarity": 0.071, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.076, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46836, + "code_similarity": 0.238, + "size_similarity": 0.715, + "file_overlap": 0.152, + "area_overlap": 0.0, + "patch_similarity": 0.33, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "src/transformers/conversion_mapping.py", + "src/transformers/models/__init__.py", + "src/transformers/models/auto/auto_mappings.py", + "src/transformers/models/auto/modeling_auto.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 1361, + 1368 + ] + ], + "right_ranges": [ + [ + 885, + 892 + ] + ] + }, + { + "filename": "src/transformers/conversion_mapping.py", + "left_ranges": [ + [ + 627, + 659 + ] + ], + "right_ranges": [ + [ + 1380, + 1446 + ] + ] + }, + { + "filename": "src/transformers/models/__init__.py", + "left_ranges": [ + [ + 288, + 294 + ] + ], + "right_ranges": [ + [ + 450, + 456 + ] + ] + }, + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 385, + 392 + ], + [ + 813, + 819 + ], + [ + 933, + 939 + ], + [ + 1054, + 1060 + ], + [ + 1186, + 1192 + ] + ], + "right_ranges": [ + [ + 623, + 633 + ], + [ + 898, + 907 + ], + [ + 1102, + 1108 + ], + [ + 1228, + 1234 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 325, + 332 + ], + [ + 1081, + 1087 + ], + [ + 1900, + 1906 + ] + ], + "right_ranges": [ + [ + 495, + 501 + ], + [ + 1097, + 1103 + ] + ] + } + ] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46850, + "code_similarity": 0.027, + "size_similarity": 0.093, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.055, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46891, + "code_similarity": 0.002, + "size_similarity": 0.007, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.007, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46953, + "code_similarity": 0.002, + "size_similarity": 0.01, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.003, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43451, + "right_pr_number": 46957, + "code_similarity": 0.012, + "size_similarity": 0.029, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.042, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46707, + "code_similarity": 0.027, + "size_similarity": 0.074, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.079, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46836, + "code_similarity": 0.205, + "size_similarity": 0.748, + "file_overlap": 0.029, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [ + "src/transformers/models/auto/auto_mappings.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/auto/auto_mappings.py", + "left_ranges": [ + [ + 362, + 368 + ], + [ + 591, + 597 + ], + [ + 796, + 802 + ] + ], + "right_ranges": [ + [ + 623, + 633 + ], + [ + 898, + 907 + ], + [ + 1102, + 1108 + ], + [ + 1228, + 1234 + ] + ] + } + ] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46850, + "code_similarity": 0.028, + "size_similarity": 0.098, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.054, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46891, + "code_similarity": 0.003, + "size_similarity": 0.007, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.008, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46953, + "code_similarity": 0.003, + "size_similarity": 0.011, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.003, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46658, + "right_pr_number": 46957, + "code_similarity": 0.013, + "size_similarity": 0.03, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.046, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46836, + "code_similarity": 0.031, + "size_similarity": 0.099, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.078, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46850, + "code_similarity": 0.161, + "size_similarity": 0.76, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.058, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46891, + "code_similarity": 0.021, + "size_similarity": 0.093, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.014, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46953, + "code_similarity": 0.031, + "size_similarity": 0.143, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.015, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46707, + "right_pr_number": 46957, + "code_similarity": 0.121, + "size_similarity": 0.403, + "file_overlap": 0.1, + "area_overlap": 0.0, + "patch_similarity": 0.069, + "shared_filenames": [ + "src/transformers/generation/utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/generation/utils.py", + "left_ranges": [ + [ + 33, + 46 + ] + ], + "right_ranges": [ + [ + 2156, + 2186 + ], + [ + 2865, + 2871 + ] + ] + } + ] + }, + { + "left_pr_number": 46836, + "right_pr_number": 46850, + "code_similarity": 0.035, + "size_similarity": 0.131, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.057, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46836, + "right_pr_number": 46891, + "code_similarity": 0.003, + "size_similarity": 0.009, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.008, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46836, + "right_pr_number": 46953, + "code_similarity": 0.003, + "size_similarity": 0.014, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.003, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46836, + "right_pr_number": 46957, + "code_similarity": 0.015, + "size_similarity": 0.04, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.049, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46850, + "right_pr_number": 46891, + "code_similarity": 0.017, + "size_similarity": 0.071, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.019, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46850, + "right_pr_number": 46953, + "code_similarity": 0.027, + "size_similarity": 0.109, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.033, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46850, + "right_pr_number": 46957, + "code_similarity": 0.071, + "size_similarity": 0.306, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.068, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46891, + "right_pr_number": 46953, + "code_similarity": 0.13, + "size_similarity": 0.652, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.0, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46891, + "right_pr_number": 46957, + "code_similarity": 0.049, + "size_similarity": 0.231, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.019, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 46953, + "right_pr_number": 46957, + "code_similarity": 0.075, + "size_similarity": 0.355, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.027, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46894-4", + "summary": "Cluster of 2 issues and 2 PRs centered on issue #46894.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 46894, + "canonical_pr_number": 46895, + "issue_numbers": [ + 46894, + 46897 + ], + "pr_numbers": [ + 46895, + 46958 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46895, + "right_pr_number": 46958, + "code_similarity": 0.715, + "size_similarity": 0.833, + "file_overlap": 1.0, + "area_overlap": 0.366, + "patch_similarity": 0.802, + "shared_filenames": [ + "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "tests/models/cohere_asr/test_modeling_cohere_asr.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/cohere_asr/modeling_cohere_asr.py", + "left_ranges": [ + [ + 615, + 627 + ], + [ + 639, + 654 + ] + ], + "right_ranges": [ + [ + 22, + 28 + ], + [ + 638, + 645 + ] + ] + }, + { + "filename": "src/transformers/models/cohere_asr/modular_cohere_asr.py", + "left_ranges": [ + [ + 478, + 490 + ], + [ + 502, + 517 + ] + ], + "right_ranges": [ + [ + 16, + 22 + ], + [ + 501, + 508 + ] + ] + }, + { + "filename": "tests/models/cohere_asr/test_modeling_cohere_asr.py", + "left_ranges": [ + [ + 35, + 41 + ], + [ + 176, + 214 + ] + ], + "right_ranges": [ + [ + 15, + 21 + ], + [ + 180, + 221 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45647-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45647.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45647, + "canonical_pr_number": 45738, + "issue_numbers": [ + 45647 + ], + "pr_numbers": [ + 45737, + 45738 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45737, + "right_pr_number": 45738, + "code_similarity": 0.526, + "size_similarity": 0.11, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.024, + "shared_filenames": [ + "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py", + "left_ranges": [ + [ + 491, + 497 + ] + ], + "right_ranges": [ + [ + 491, + 497 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46869-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46869.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 46869, + "canonical_pr_number": 46886, + "issue_numbers": [ + 46869 + ], + "pr_numbers": [ + 46886, + 46922 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 46886, + "right_pr_number": 46922, + "code_similarity": 0.711, + "size_similarity": 0.857, + "file_overlap": 1.0, + "area_overlap": 0.5, + "patch_similarity": 0.429, + "shared_filenames": [ + "src/transformers/models/yolos/modeling_yolos.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/yolos/modeling_yolos.py", + "left_ranges": [ + [ + 446, + 452 + ] + ], + "right_ranges": [ + [ + 434, + 440 + ], + [ + 446, + 452 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46159-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46159.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 46159, + "canonical_pr_number": 46174, + "issue_numbers": [ + 46159 + ], + "pr_numbers": [ + 46174, + 46780 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 46174, + "right_pr_number": 46780, + "code_similarity": 0.841, + "size_similarity": 0.87, + "file_overlap": 1.0, + "area_overlap": 0.767, + "patch_similarity": 0.657, + "shared_filenames": [ + "src/transformers/models/aimv2/modeling_aimv2.py", + "src/transformers/models/d_fine/modeling_d_fine.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "src/transformers/models/pp_doclayout_v2/modeling_pp_doclayout_v2.py", + "src/transformers/models/pp_doclayout_v3/modeling_pp_doclayout_v3.py", + "src/transformers/models/rt_detr/modeling_rt_detr.py", + "src/transformers/models/rt_detr_v2/modeling_rt_detr_v2.py", + "src/transformers/models/vit_mae/modeling_vit_mae.py", + "src/transformers/models/vit_mae/modular_vit_mae.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/aimv2/modeling_aimv2.py", + "left_ranges": [ + [ + 134, + 141 + ], + [ + 144, + 154 + ], + [ + 157, + 165 + ] + ], + "right_ranges": [ + [ + 134, + 140 + ], + [ + 143, + 153 + ], + [ + 156, + 162 + ] + ] + }, + { + "filename": "src/transformers/models/d_fine/modeling_d_fine.py", + "left_ranges": [ + [ + 631, + 638 + ], + [ + 641, + 651 + ], + [ + 654, + 662 + ] + ], + "right_ranges": [ + [ + 631, + 637 + ], + [ + 640, + 650 + ], + [ + 653, + 659 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/modeling_deimv2.py", + "left_ranges": [ + [ + 703, + 710 + ], + [ + 713, + 723 + ], + [ + 726, + 734 + ] + ], + "right_ranges": [ + [ + 703, + 709 + ], + [ + 712, + 722 + ], + [ + 725, + 731 + ] + ] + }, + { + "filename": "src/transformers/models/pp_doclayout_v2/modeling_pp_doclayout_v2.py", + "left_ranges": [ + [ + 1551, + 1558 + ], + [ + 1561, + 1571 + ], + [ + 1574, + 1582 + ] + ], + "right_ranges": [ + [ + 1551, + 1557 + ], + [ + 1560, + 1570 + ], + [ + 1573, + 1579 + ] + ] + }, + { + "filename": "src/transformers/models/pp_doclayout_v3/modeling_pp_doclayout_v3.py", + "left_ranges": [ + [ + 804, + 811 + ], + [ + 814, + 824 + ], + [ + 827, + 835 + ] + ], + "right_ranges": [ + [ + 804, + 810 + ], + [ + 813, + 823 + ], + [ + 826, + 832 + ] + ] + }, + { + "filename": "src/transformers/models/rt_detr/modeling_rt_detr.py", + "left_ranges": [ + [ + 847, + 854 + ], + [ + 857, + 867 + ], + [ + 870, + 878 + ] + ], + "right_ranges": [ + [ + 847, + 853 + ], + [ + 856, + 866 + ], + [ + 869, + 875 + ] + ] + }, + { + "filename": "src/transformers/models/rt_detr_v2/modeling_rt_detr_v2.py", + "left_ranges": [ + [ + 976, + 983 + ], + [ + 986, + 996 + ], + [ + 999, + 1007 + ] + ], + "right_ranges": [ + [ + 976, + 982 + ], + [ + 985, + 995 + ], + [ + 998, + 1004 + ] + ] + }, + { + "filename": "src/transformers/models/vit_mae/modeling_vit_mae.py", + "left_ranges": [ + [ + 152, + 159 + ], + [ + 162, + 172 + ], + [ + 175, + 183 + ] + ], + "right_ranges": [ + [ + 152, + 158 + ], + [ + 161, + 171 + ], + [ + 174, + 180 + ] + ] + }, + { + "filename": "src/transformers/models/vit_mae/modular_vit_mae.py", + "left_ranges": [ + [ + 62, + 69 + ], + [ + 72, + 82 + ], + [ + 85, + 93 + ] + ], + "right_ranges": [ + [ + 62, + 68 + ], + [ + 71, + 81 + ], + [ + 84, + 90 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.721, + "size_similarity": 0.973, + "file_overlap": 0.75, + "area_overlap": 0.563, + "patch_similarity": 0.696, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py", + "utils/notification_service.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3611, + 3618 + ], + [ + 3835, + 3869 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 144 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + }, + { + "filename": "utils/notification_service.py", + "left_ranges": [ + [ + 935, + 967 + ] + ], + "right_ranges": [ + [ + 935, + 967 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-46305-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #46305.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 46305, + "canonical_pr_number": 46546, + "issue_numbers": [ + 46305 + ], + "pr_numbers": [ + 46310, + 46546 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 46310, + "right_pr_number": 46546, + "code_similarity": 0.566, + "size_similarity": 0.444, + "file_overlap": 0.5, + "area_overlap": 0.867, + "patch_similarity": 0.16, + "shared_filenames": [ + "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/moonshine_streaming/modeling_moonshine_streaming.py", + "left_ranges": [ + [ + 356, + 362 + ] + ], + "right_ranges": [ + [ + 355, + 361 + ] + ] + }, + { + "filename": "src/transformers/models/moonshine_streaming/modular_moonshine_streaming.py", + "left_ranges": [ + [ + 269, + 275 + ] + ], + "right_ranges": [ + [ + 269, + 275 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, + "file_overlap": 1.0, + "area_overlap": 0.425, + "patch_similarity": 0.866, + "shared_filenames": [ + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "left_ranges": [ + [ + 26, + 46 + ], + [ + 138, + 143 + ], + [ + 182, + 187 + ], + [ + 283, + 295 + ], + [ + 327, + 341 + ], + [ + 348, + 354 + ], + [ + 360, + 369 + ], + [ + 399, + 406 + ], + [ + 411, + 419 + ], + [ + 428, + 433 + ], + [ + 472, + 492 + ], + [ + 507, + 513 + ], + [ + 519, + 528 + ], + [ + 541, + 559 + ], + [ + 562, + 567 + ], + [ + 595, + 601 + ], + [ + 607, + 614 + ], + [ + 627, + 643 + ], + [ + 685, + 690 + ], + [ + 708, + 714 + ], + [ + 720, + 727 + ], + [ + 740, + 757 + ], + [ + 761, + 766 + ], + [ + 780, + 786 + ], + [ + 791, + 798 + ], + [ + 807, + 822 + ], + [ + 840, + 845 + ] + ], + "right_ranges": [ + [ + 26, + 31 + ], + [ + 34, + 44 + ], + [ + 136, + 141 + ], + [ + 180, + 185 + ], + [ + 281, + 293 + ], + [ + 325, + 339 + ], + [ + 346, + 352 + ], + [ + 358, + 367 + ], + [ + 397, + 404 + ], + [ + 409, + 417 + ], + [ + 426, + 434 + ], + [ + 463, + 483 + ], + [ + 498, + 504 + ], + [ + 510, + 519 + ], + [ + 532, + 550 + ], + [ + 553, + 564 + ], + [ + 586, + 592 + ], + [ + 598, + 605 + ], + [ + 618, + 634 + ], + [ + 676, + 688 + ], + [ + 699, + 705 + ], + [ + 711, + 718 + ], + [ + 731, + 748 + ], + [ + 752, + 762 + ], + [ + 771, + 777 + ], + [ + 782, + 789 + ], + [ + 798, + 813 + ], + [ + 831, + 836 + ] + ] + }, + { + "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", + "left_ranges": [ + [ + 458, + 464 + ] + ], + "right_ranges": [ + [ + 458, + 464 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [ + { + "cluster_id": "cluster-46894-4", + "canonical_issue_number": 46894, + "duplicate_issue_numbers": [ + 46897 + ], + "reason": "Issues in cluster-46894-4 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." + } + ], + "duplicate_prs": [ + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43451-8", + "canonical_pr_number": 43451, + "duplicate_pr_numbers": [ + 46658, + 46707, + 46836, + 46850, + 46891, + 46953, + 46957 + ], + "target_issue_number": null, + "reason": "PRs in cluster-43451-8 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46159-3", + "canonical_pr_number": 46174, + "duplicate_pr_numbers": [ + 46780 + ], + "target_issue_number": 46159, + "reason": "PRs in cluster-46159-3 are treated as duplicates because they converge on issue #46159 with closing_reference evidence." + }, + { + "cluster_id": "cluster-46305-3", + "canonical_pr_number": 46546, + "duplicate_pr_numbers": [ + 46310 + ], + "target_issue_number": 46305, + "reason": "PRs in cluster-46305-3 are treated as duplicates because they converge on issue #46305 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-46869-3", + "canonical_pr_number": 46886, + "duplicate_pr_numbers": [ + 46922 + ], + "target_issue_number": 46869, + "reason": "PRs in cluster-46869-3 are treated as duplicates because they converge on issue #46869 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-46894-4", + "canonical_pr_number": 46895, + "duplicate_pr_numbers": [ + 46958 + ], + "target_issue_number": 46894, + "reason": "PRs in cluster-46894-4 are treated as duplicates because they converge on issue #46894 with closing_reference, shared_issue_target, soft_similarity evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-46894-4", + "issue_number": 46894, + "reason": "Issue #46894 is the strongest global issue candidate because it is open, belongs to a cluster with 4 artifacts, and carries 0 discussion comments plus 6 inbound references.", + "score": 146.0 + }, + "best_pr": { + "cluster_id": "cluster-43979-11", + "pr_number": 44007, + "reason": "PR #44007 is the strongest global PR candidate because it is open, belongs to a cluster with 11 artifacts, links to 1 issues, and carries 2 review/discussion events.", + "score": 238.886 + } +} diff --git a/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..118778e3a235272cc0aa35d9d0e2b115c7b62c32 --- /dev/null +++ b/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,25825 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260630T000016Z", + "generated_at": "2026-06-30T00:09:40Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43425", + "issue:43541", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43827", + "issue:43828", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44841", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:46015", + "issue:46165", + "issue:46294", + "issue:46468", + "issue:46515", + "issue:46661", + "issue:46869" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:46468", + "issue:45237|issue:45507", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44918|issue:46165", + "issue:46165|issue:46869", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:46294", + "issue:44805|issue:45161", + "issue:46165|issue:46468", + "issue:44560|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:44841|issue:46661", + "issue:43716|issue:46869", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:45685|issue:46869", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:45874|issue:46294", + "issue:44292|issue:45237", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:45800|issue:46515", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:43828|issue:45161", + "issue:43828|issue:44292", + "issue:43541|issue:43716", + "issue:43716|issue:46468", + "issue:44292|issue:44918", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:44918|issue:45161", + "issue:43716|issue:45685", + "issue:41762|issue:43638", + "issue:43039|issue:44918", + "issue:43827|issue:44512", + "issue:43828|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:45237|issue:46015", + "issue:43039|issue:44805" + ], + "prepared_review_unit_hash": "47580346ddada2b540346b57f447a3c240eb25310bae3ce3ca62120f5963ddc6", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 63, + "serialized_chars": 46571, + "estimated_input_tokens": 11643, + "estimated_eval_tokens": 23542 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: a few true duplicates around ZeRO-3 / torch-version / Qwen3.5-MoE rotary-embedding issues, but many other pairs are only loosely related model/runtime failures or docs regressions and should stay separate.", + "confidence": 0.56, + "canonical_issue_reason": "Issue 45237 is the best cluster representative because it sits in the densest subgroup of model/device-placement failures (GPU, auto device map, offload, MoE/runtime errors) and is broad enough to summarize the dominant theme.", + "canonical_pr_reason": null, + "best_issue_reason": "45237 is the most suitable issue to keep as the global representative: it is concrete, central to the cluster\u2019s main runtime/device-mapping theme, and broader than the more model-specific outliers.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: Qwen3.5 config attribute vs streaming chat/completions API." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different concrete bugs: AMD GPU failure vs MPS histogram kernel limitation." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Both are accelerator-related, but they affect different model paths and error types." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is a different code path from GPT-OSS AMD GPU failure." + }, + { + "left": "issue:45237", + "right": "issue:46468", + "accept": false, + "reason": "Different models and symptoms; no shared concrete bug." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "GPU/runtime failure vs invalid method call in GraniteMoEHybrid; not the same bug." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch is unrelated to ONNX export." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ModernBERT ONNX export and Qwen3Moe FineGrainedFP8 are unrelated." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export error vs Qwen3 auto device-map assert are different problems." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Table QA crash and RAG docs cleanup are unrelated." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs regressions, but they reference different removed v5 features and are not the same issue." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export failure is unrelated to GenerationMixin import breakage." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "TRL SFT embedding unpacking vs resume-training with remote code are different failures." + }, + { + "left": "issue:46165", + "right": "issue:46869", + "accept": false, + "reason": "Resume-training issue is unrelated to YOLO multi-GPU device-map failure." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Mask/tensor shape mismatch is not the same as Gemma4 PLE device mismatch." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are IndexErrors, but the tensor shape mismatch is a different failure mode from the ZeRO-3 empty-index bug." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "ZeRO-3 Bert index error is unrelated to resume-training with local models." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Similar IndexError wording, but the underlying bugs are different." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Mask shape mismatch is unrelated to Qwen3.5 video token/feature mismatch." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Different models and different device/matrix errors." + }, + { + "left": "issue:46165", + "right": "issue:46468", + "accept": false, + "reason": "Training resume problem is unrelated to Nemotron example breakage." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video input StopIteration and training resume are unrelated." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice_preset bug is unrelated to ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed, but they are different concrete failures and code paths." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same ZeRO-3/Bert empty-index failure; titles differ only by model phrasing." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross_entropy call and GPT-OSS TP failure are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Different problems: detokenization bug vs LoRA merge/save mismatch." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen3-8B NVFP4 runtime error is not the same as LoRA save/reload inconsistency." + }, + { + "left": "issue:44841", + "right": "issue:46661", + "accept": false, + "reason": "Processor failure for Voxtral is unrelated to remote-code model_class config_class handling." + }, + { + "left": "issue:43716", + "right": "issue:46869", + "accept": false, + "reason": "Different models and failure modes; only broad device issues overlap." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "FSDP checkpointing recompute mismatch is not the same as the mask/tensor shape IndexError." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "DeepSpeed weight-dimension error and collective mismatch are distinct bugs." + }, + { + "left": "issue:45685", + "right": "issue:46869", + "accept": false, + "reason": "MPS histogram issue is unrelated to YOLO multi-GPU device-map auto." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": true, + "reason": "Same Qwen3_5MoeTextRotaryEmbedding offload/materialization bug surface; one report emphasizes CPU offload, the other uninitialized inv_freq state." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "Both are model/runtime issues, but Bert ZeRO-3 index error and GPT-OSS TP failure are not the same bug." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "Activation-checkpointing/FSDP mismatch is unrelated to the CustomTrainer compute_loss signature change." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "ZeRO-3 Bert index error is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Mistral-3 dtype mismatch and GraniteMoEHybrid invalid method are different issues." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Gemma4 reverse-embedding lookup bug is unrelated to Qwen3.5 video token mismatch." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Qwen3-8B NVFP4 runtime error and GPT-OSS AMD GPU failure are different model/runtime bugs." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3 sliding_window_pattern/vLLM regression is unrelated to Mistral regex patch crash." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Gemma4 PLE device mismatch and reverse embedding lookup are separate bugs." + }, + { + "left": "issue:45800", + "right": "issue:46515", + "accept": true, + "reason": "Same torch 2.4.1 incompatibility report, just phrased differently." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch version incompatibility and Qwen3 NVFP4 runtime error are not the same issue." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP activation-checkpointing mismatch is unrelated to Qwen3 auto device-map assert." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Autocast dtype mismatch is unrelated to the MPS histogram kernel limitation." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral grouped_mm torch-dynamo tracing error is unrelated to Mistral-3 dtype mismatch." + }, + { + "left": "issue:43716", + "right": "issue:46468", + "accept": false, + "reason": "Different models and failure modes; not the same underlying bug." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Qwen3 NVFP4 runtime error and TRL SFT embedding unpacking are distinct." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call is unrelated to MPS Int histogram support." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different models and errors; only broad runtime similarity." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL embedding unpacking failure is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Both are ZeRO-3 IndexErrors, but they involve different models and likely different root causes." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger Kernel cross_entropy call and Qwen3.5 SFT trainer error are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs regressions, but they concern different removed v5 topics and should not be merged as one bug." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Phi autocast dtype mismatch and MPS histogram error are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross_entropy issue and mask/index shape mismatch are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the cluster as mostly heterogeneous, while identifying the few clearly overlapping pairs (43596/43638, 45859/45902, 45800/46515). The negative soft-edge calls are generally conservative and supported by the titles, and the choice of 45237 as the broad representative is reasonable for the device/runtime theme." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 2, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:36010", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:43039", + "issue:43296", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43792", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44512", + "issue:44560", + "issue:44589", + "issue:44841", + "issue:44863", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45237", + "issue:45362", + "issue:45399", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304", + "issue:46402", + "issue:46661", + "issue:46772", + "issue:46869" + ], + "soft_pairs": [ + "issue:43866|issue:46402", + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:41762|issue:43596", + "issue:39290|issue:41720", + "issue:43901|issue:44512", + "issue:44589|issue:46304", + "issue:45237|issue:45901", + "issue:43975|issue:46015", + "issue:43039|issue:44560", + "issue:45823|issue:46294", + "issue:44841|issue:45823", + "issue:43866|issue:46772", + "issue:44863|issue:46015", + "issue:45020|issue:46304", + "issue:45812|issue:45920", + "issue:43866|issue:44863", + "issue:44292|issue:46015", + "issue:43872|issue:46304", + "issue:39290|issue:43531", + "issue:41093|issue:43638", + "issue:43792|issue:45901", + "issue:43854|issue:43866", + "issue:43425|issue:45070", + "issue:43975|issue:45901", + "issue:45362|issue:45464", + "issue:43596|issue:45800", + "issue:43792|issue:45070", + "issue:43541|issue:46869", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:36010|issue:39290", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:44841|issue:45084", + "issue:43541|issue:43828", + "issue:43404|issue:43866", + "issue:45507|issue:46869", + "issue:41093|issue:41762", + "issue:45399|issue:46661", + "issue:43572|issue:45070", + "issue:45800|issue:45823" + ], + "prepared_review_unit_hash": "47bcb7356cf97a5a065b79aa2e0846ce1cf5f656fea68f0f637670c817fd8053", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 42, + "serialized_chars": 40038, + "estimated_input_tokens": 10010, + "estimated_eval_tokens": 20276 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous, but two soft edges look like true duplicate subsets: the DeepSpeed ZeRO-3 load-time IndexError reports and the v5 AutoTokenizer wrong-token-id regression reports.", + "confidence": 0.61, + "canonical_issue_reason": "issue:45920 is the best canonical issue for the tokenizer-regression subset because it describes the broad v5 AutoTokenizer wrong-token-ID bug across multiple model families, subsuming the narrower Granite-only report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the strongest overall issue anchor: it is broad, specific about the regression, and clearly actionable.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Different models and different failure modes: corrupted checkpoint vs. missing weight loading in Sam3Model." + }, + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "Both are compatibility complaints, but they point to different breaks and different trigger conditions." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM is unrelated to the USE_DISTRIBUTED=0 PyTorch support regression." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Both report the same ZeRO-3 load/init path failing with an index-0 out-of-bounds error; model names differ, but the underlying bug appears shared." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3 config attribute error and Qwen3 auto-device-map cuda assert are different code paths and symptoms." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues, but they concern different API docs and do not share the same underlying bug." + }, + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to the PyTorch distributed support regression." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "AMD GPU support for GPT-OSS is unrelated to the table-question-answering crash." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Incorrect detokenization and outdated RAG examples are unrelated issues." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross-entropy behavior and Qwen3-vl video StopIteration are different problems." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Gemma4 device-mismatch on auto device mapping is unrelated to Qwen3.5 video token/template mismatches." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Voxtral processor failure and Gemma4 device mismatch are distinct model-specific issues." + }, + { + "left": "issue:43866", + "right": "issue:46772", + "accept": false, + "reason": "Checkpoint corruption and NVFP4/ModelOpt compatibility are not the same bug." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint loading and RAG docs are unrelated." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Remote-code breakage and PyTorch USE_DISTRIBUTED=0 support are different regressions." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "These are the same v5 AutoTokenizer wrong-token-ID regression; 45920 is the broader report covering Granite and other affected model families." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted checkpoint vs. checkpoint loading failure for a different model family are not the same underlying bug." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen NVFP4 runtime failure is unrelated to RAG documentation examples." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes Int8Params incompatibility is unrelated to the distributed-build regression." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but one is a missing config attribute in Gemma3/vLLM while the other is a Qwen3-MoE sliding_window issue; too different to merge." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "The mask-shape mismatch and the ZeRO-3 empty-index error are different IndexError classes with different contexts." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper model loading and table-question-answering crashes are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "GLM-4.7-Flash loading in tests and Ovis2 checkpoint corruption are distinct model-specific issues." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch-version incompatibility and pydantic PretrainedConfig breakage are unrelated regressions." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Detokenization bug and table-question-answering crash are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both involve Qwen3.5 chat, but one is a transformers chat crash and the other is a streaming chat/completions API failure; not the same concrete bug." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "ZeRO-3 index error and torch/transformers version incompatibility are different problems." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper loading and pydantic model breakage are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:46869", + "accept": false, + "reason": "Mixtral grouped_mm tracing failure and YOLO multi-GPU device-map failure are different issues." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "NVFP4 runtime failure and table QA crash are unrelated." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3 sliding_window_pattern attribute error and Gemma4 device-mismatch are different model/path failures." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "GenerationMixin import error and Gemma3TextConfig sliding_window_pattern error are unrelated." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3.5 streaming inference and MPS histogram_int support are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Mixtral tracing failure and MPS histogram_int error are unrelated." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Voxtral processor failure and template-node compilation error are different bugs." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different MoE-related symptoms, not the same code-path bug." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Mistral3 lm_head tying issue and Ovis2 checkpoint corruption are unrelated." + }, + { + "left": "issue:45507", + "right": "issue:46869", + "accept": false, + "reason": "GraniteMoEHybrid invalid method and YOLO multi-GPU auto device-map failure are different model-specific issues." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are IndexErrors, but the mask/tensor shape mismatch is not the same as the ZeRO-3 empty-index load failure." + }, + { + "left": "issue:45399", + "right": "issue:46661", + "accept": false, + "reason": "Flash-attn fallback blocking and remote-code config_class errors are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig and pydantic PretrainedConfig breakage are different regressions." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers compatibility and Gemma4 device mismatch are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The two accepted soft edges are well-supported by the titles: the ZeRO-3 index-out-of-bounds reports share the same failure shape, and 45920 clearly subsumes the Granite-only AutoTokenizer regression in 45812. No evidence looks overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 3, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:39186", + "issue:40990", + "issue:41093", + "issue:42915", + "issue:43054", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43596", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43873", + "issue:43994", + "issue:44291", + "issue:44360", + "issue:44387", + "issue:44479", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44805", + "issue:44863", + "issue:44918", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45161", + "issue:45183", + "issue:45399", + "issue:45464", + "issue:45632", + "issue:45698", + "issue:45759", + "issue:45800", + "issue:45902", + "issue:46165", + "issue:46294", + "issue:46463", + "issue:46468", + "issue:46489", + "issue:46491", + "issue:46640", + "issue:46661", + "issue:46869" + ], + "soft_pairs": [ + "issue:43296|issue:43366", + "issue:44805|issue:44918", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:43366|issue:45464", + "issue:44291|issue:45070", + "issue:43531|issue:45070", + "issue:41093|issue:43596", + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:43054|issue:43994", + "issue:44479|issue:46294", + "issue:46468|issue:46869", + "issue:45759|issue:45902", + "issue:43596|issue:46165", + "issue:46640|issue:46661", + "issue:43716|issue:44918", + "issue:39186|issue:44918", + "issue:40990|issue:44960", + "issue:43645|issue:45399", + "issue:45632|issue:45698", + "issue:43404|issue:45005", + "issue:42915|issue:45399", + "issue:43873|issue:45005", + "issue:44530|issue:45005", + "issue:44360|issue:44512", + "issue:44492|issue:44512", + "issue:44863|issue:45005", + "issue:46489|issue:46491", + "issue:45183|issue:46463" + ], + "prepared_review_unit_hash": "7053eab7b8448d871516f63d8695caa83f41a47f94de8eebad75a15bf3777738", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36891, + "estimated_input_tokens": 9223, + "estimated_eval_tokens": 18702 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous bucket of unrelated issues; only a few soft edges look like real duplicates. The strongest duplicate-like groups are the Qwen3.5 video-input regression, the remote-code/custom-module loading collision, and the input validation error-message bug. Most other pairs are different model families or different failure modes.", + "confidence": 0.74, + "canonical_issue_reason": "issue:45183 is the broadest root-cause report in the set and cleanly subsumes the T5-specific validation symptom in issue:46463.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45183 is the best standalone representative because it describes the underlying validation bug, not just one model-family manifestation.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "Different problem domains: PaddleOCR-VL load failure vs gpt-oss GGUF architecture support." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Both mention shape-related failures, but the code paths and symptoms differ (mask indexing vs TRL embedding unpacking)." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Unrelated: int4 quantization memory/OOM vs tied-weights behavior on translation models." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Both involve MoE models, but one is a grouped_mm tracing error and the other is a tensor-parallel support issue." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: autocast dtype mismatch vs chat/completions streaming inference on Qwen3.5-0.8B." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "No concrete overlap beyond both being Qwen/gpt-oss-era model issues." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Both are v5 regressions, but one is init_empty_weights argument handling and the other is a PretrainedConfig/pydantic field breakage." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Different bug classes: sliding_window handling vs config serialization/type issues." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Mask-length mismatch and zero-sized index out-of-bounds are distinct runtime errors in different paths." + }, + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Both are multimodal input mismatches, but image preprocessor dtype mismatch and video token/feature mismatch are not the same bug." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "A broad remote_code regression and a torch/transformers version incompatibility are different root causes." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but one is a quality/performance complaint while the other is an inference correctness bug; not clearly the same code path." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": true, + "reason": "46294 looks like a specific Qwen3.5 video-input manifestation of the broader video regression described in 44479." + }, + { + "left": "issue:46468", + "right": "issue:46869", + "accept": false, + "reason": "Official Nemotron example breakage and YOLO multi-GPU device-map failure are unrelated." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Different model/config problems: composite config unwrapping vs uninitialized rotary embedding memory after materialization." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "Different areas and symptoms: Deepspeed zero3/BertModel index error vs trust_remote_code resume behavior." + }, + { + "left": "issue:46640", + "right": "issue:46661", + "accept": false, + "reason": "Flash Attention 2 fallback logic and remote-code model_class config_class errors are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Different multimodal/text paths; image preprocessor dtype mismatch is not the same as TRL embedding unpacking failure." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP weight dimensionality error and Qwen3.5 embedding unpacking are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and GLM5 are separate model-specific reports." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Custom-model notebook initialization and flash-attn2 fallback gating are different bugs." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": true, + "reason": "Both point to remote-code/custom-module loading selecting the wrong local module; the latter is a concrete symptom of the cache/path collision problem." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Tied lm_head weights in Mistral3 and translation-model tied-weights behavior are different model families and failures." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "FineGrainedFP8Config failure and flash-attn2 fallback blocking are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior and tied-weight translation issues are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash is not the same as tied-weights handling." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "DSA indexer/ReLU discussion and docs for removed run command are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A documentation typo and a removed-command docs issue are not duplicates." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading and tied weights in translation models are different bugs." + }, + { + "left": "issue:46489", + "right": "issue:46491", + "accept": false, + "reason": "Different tokenizer regressions affecting different models and different round-trip behaviors." + }, + { + "left": "issue:45183", + "right": "issue:46463", + "accept": true, + "reason": "46463 is a T5-family-specific instance of the broader input_ids/inputs_embeds validation/error-message bug in 45183." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately cautious. The three accepted soft edges are plausible duplicates/specific manifestations from the titles and descriptions, and the rejected pairs are conservatively framed as different bug classes or model families. I don\u2019t see an overstatement that would require rejection." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 4, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:24643", + "issue:28530", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41553", + "issue:41762", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43493", + "issue:43643", + "issue:43644", + "issue:43782", + "issue:43824", + "issue:43828", + "issue:43856", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44292", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44589", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45399", + "issue:45507", + "issue:45520", + "issue:45663", + "issue:45701", + "issue:45859", + "issue:45902", + "issue:45920", + "issue:45987", + "issue:46489", + "issue:46858", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:43975|issue:44292", + "issue:44918|issue:45507", + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue:36331|issue:41762", + "issue:43872|issue:44589", + "issue:46899|issue:46932", + "issue:43828|issue:44928", + "issue:45920|issue:46489", + "issue:45987|issue:46858", + "issue:43054|issue:45701", + "issue:38175|issue:43054", + "issue:43950|issue:45902", + "issue:41553|issue:43883", + "issue:43054|issue:43493", + "issue:44910|issue:45859", + "issue:42947|issue:44387", + "issue:45663|issue:45902", + "issue:44403|issue:44451", + "issue:43975|issue:45520", + "issue:41553|issue:43881", + "issue:24643|issue:28530", + "issue:42947|issue:43856", + "issue:43643|issue:45020", + "issue:44403|issue:44488", + "issue:41762|issue:42915", + "issue:44991|issue:45399", + "issue:43782|issue:44960", + "issue:43824|issue:43872" + ], + "prepared_review_unit_hash": "a7db8f174bdefa87da636906d175869ca00b0c36e2d6ff6eb9b58c0fa6139798", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36789, + "estimated_input_tokens": 9198, + "estimated_eval_tokens": 18652 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, with a few true duplicate chains around specific regressions. The strongest duplicate set is the Transformers 5.x non-persistent-buffer regression, plus a separate exact match around Gemma4 quantized embedders and another around cache sequence-length return types.", + "confidence": 0.78, + "canonical_issue_reason": "issue:43950 is the best umbrella issue: it states the broad non-persistent-buffer regression clearly and covers the closely related manifestations in issue:44534 and issue:45902.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 is the most representative standalone issue in the cluster because it captures the shared underlying bug rather than one narrow symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Both are model/tokenizer loading failures, but they involve different models and different failure modes; not the same bug." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Unrelated problems: one is TRL SFT unpacking, the other is an invalid method call in a different model class." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying v5 non-persistent-buffer regression; 44534 is essentially the same symptom as 43950." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP2, but one is a documentation/example mismatch and the other is a quality discrepancy; not the same concrete bug." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": false, + "reason": "Different areas of the codebase: remote-code config loading vs buffer materialization regression." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Different training/loading failures with no shared code path indicated." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different quantization/storage incompatibilities; the titles do not indicate the same underlying fix." + }, + { + "left": "issue:46899", + "right": "issue:46932", + "accept": true, + "reason": "46932 explicitly describes the same quantized vision/audio dtype-casting bug as 46899." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Both are numerical/runtime issues, but they are different models and different mechanisms; not a duplicate." + }, + { + "left": "issue:45920", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but the titles point to different affected models and a broader vs specific failure; not enough to treat as the same bug." + }, + { + "left": "issue:45987", + "right": "issue:46858", + "accept": true, + "reason": "Same cache-length contract bug: get_seq_length is returning a tensor instead of an int/device scalar, breaking downstream arithmetic." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "One is a SigLIP2 embedding-quality complaint, the other is a broad tokenization-change report; unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both concern SigLIP2 outputs, but one is zero probabilities and the other is lower text embedding quality; not clearly the same root cause." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": true, + "reason": "Both describe non-persistent buffers becoming junk/uninitialized after loading or materialization in Transformers 5.x." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Different model-loading errors with no shared failure mode." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Both are SigLIP2-related, but one is a generic performance comparison while the other is a specific HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Different Qwen3.5 issues: FlashAttention misinterpreting 3D position_ids vs CPU-offload incompatibility in rotary embeddings." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory, but one is gradient checkpointing/LoRA behavior and the other is int4 quantization OOM; different bugs." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states KeyError is unrelated to the non-persistent-buffer regression." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "One is a noisy-loading UX issue, the other is a model load failure; not duplicates." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "DeepSeek-Coder detokenization and a flash_attn import error are unrelated." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different model-loading failures and different components." + }, + { + "left": "issue:24643", + "right": "issue:28530", + "accept": false, + "reason": "Training with DeepSpeed and early-stopping metric handling are unrelated issues." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different memory-related regressions; no shared code path implied by the titles." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Both mention remote code, but one is missing fields in AutoConfig and the other is a broad breakage report; too little to merge as the same bug." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Different symptoms and likely different model-loading paths." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Completely different models and failures: ZeRO-3 loading vs FineGrainedFP8Config." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading and flash-attn fallback are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different model families and different error classes." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Import/export symbol availability and bitsandbytes parameter incompatibility are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative on the soft-edge pairs. The cluster summary correctly frames the set as mostly heterogeneous with a few duplicate chains, and the selected canonical issue/pairing around issue:43950 is well supported by the titles. I did not see any overstatement that would require rejection." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 5, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:36331", + "issue:41093", + "issue:41553", + "issue:42491", + "issue:42947", + "issue:43257", + "issue:43404", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43940", + "issue:43950", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44560", + "issue:44661", + "issue:44805", + "issue:44863", + "issue:44928", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45305", + "issue:45313", + "issue:45356", + "issue:45399", + "issue:45488", + "issue:45588", + "issue:45685", + "issue:45715", + "issue:45759", + "issue:45823", + "issue:45923", + "issue:46315", + "issue:46489", + "issue:46661", + "issue:46869", + "issue:46894", + "issue:46897" + ], + "soft_pairs": [ + "issue:45685|issue:45823", + "issue:43881|issue:43883", + "issue:45081|issue:45399", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:42491|issue:44368", + "issue:43828|issue:45823", + "issue:45020|issue:45356", + "issue:45715|issue:46315", + "issue:45488|issue:46489", + "issue:43827|issue:44509", + "issue:46894|issue:46897", + "issue:45759|issue:46661", + "issue:43404|issue:43950", + "issue:43872|issue:46869", + "issue:43872|issue:44291", + "issue:45020|issue:46661", + "issue:43792|issue:44220", + "issue:43257|issue:45305", + "issue:42947|issue:45923", + "issue:36331|issue:41093", + "issue:41553|issue:44991", + "issue:43854|issue:44863", + "issue:43940|issue:45313", + "issue:44928|issue:45588", + "issue:41093|issue:44805", + "issue:44479|issue:44560", + "issue:41553|issue:44488" + ], + "prepared_review_unit_hash": "14ba2c3c840d4f85ff4d31a021e1c9297de5624b7eb005f72d68e8f99f2af5ef", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 36193, + "estimated_input_tokens": 9049, + "estimated_eval_tokens": 18354 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous, with only a few true duplicate pockets: v5 docs still referencing removed pipeline tasks, repeated mask/tensor shape-mismatch reports, and a likely repeated training-loss double-shift report. The rest share broad subsystems or model families but do not look like the same underlying bug.", + "confidence": 0.76, + "canonical_issue_reason": "No single canonical issue cleanly covers the whole set; the items split into multiple unrelated bug pockets, so a global canonical would be misleading.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44509 is the clearest representative of the docs duplicate pocket because it explicitly names the removed v5 pipeline tasks and has the broadest, most actionable description.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "Different failures: MPS int histogram support vs device-mismatch on PLE forward." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different models and different loading bugs; only the broad symptom is similar." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Unrelated tokenizer regex crash vs flash-attn fallback gating." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Both are model-loading regressions, but the concrete failure modes are different." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Different code paths: tokenizer mapping edge case vs remote-code loading regression." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Same model family, but one is a LoRA compatibility break and the other is a warning about tie_word_embeddings." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Different runtime error classes: autocast dtype mismatch vs device placement mismatch." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer/model regressions, but not the same bug or fix." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance issues, but they hit different conversion directions and code paths." + }, + { + "left": "issue:45488", + "right": "issue:46489", + "accept": false, + "reason": "Related tokenizer regressions, but different model families and not enough evidence of one shared bug." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs regression: v5 removed pipeline tasks, and both issues report stale docs referencing them." + }, + { + "left": "issue:46894", + "right": "issue:46897", + "accept": true, + "reason": "Same training-loss double-shift pattern reported for different models; likely the same underlying loss bug." + }, + { + "left": "issue:45759", + "right": "issue:46661", + "accept": false, + "reason": "Different remote-code/composite-config failures with different missing attributes." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Different problems: untied lm_head weights vs non-persistent buffer corruption." + }, + { + "left": "issue:43872", + "right": "issue:46869", + "accept": false, + "reason": "Different device/loading contexts; bitsandbytes init arg mismatch is not the YOLO multi-GPU issue." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": false, + "reason": "Same unexpected-argument theme, but different libraries and loading paths." + }, + { + "left": "issue:45020", + "right": "issue:46661", + "accept": false, + "reason": "Both involve remote-code models, but the reported AttributeError differs from the broader v5 loading breakage." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Both involve audio models, but one is a general Whisper run failure and the other is a specific fbank-feature issue." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Different DeepSpeed problems: weight conversion/initialization vs gradient accumulation averaging." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Different checkpointing issues: ineffective gradient checkpointing vs missing capability flag." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Completely different error classes: trainer API keyword mismatch vs tensor mask shape mismatch." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/model loading problems, but for different models and different root causes." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model-loading failures for different architectures; no shared bug evidence." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Same subsystem, but different Qwen variants and different missing-weight patterns." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Different failures: NaN/SDPA math fallback vs flash-attn sink auxiliary crash." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same mask-shape mismatch error signature; these look like duplicate reports of the same indexing bug." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-related regressions, but one is a broad v5 video-input break and the other is a StopIteration embedding error." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Different tokenizer-loading issues for different models and failure modes." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model families and different checkpoint/loading failures." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 load failures, but the concrete affected models and missing-weight patterns differ." + }, + { + "left": "issue:45305", + "right": "issue:46894", + "accept": false, + "reason": "Different training bugs: GAS averaging under DeepSpeed vs a double-shift loss issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative about the few duplicate pockets it identifies. The accepted soft-edge pairs are well supported by the titles/descriptions, and the rest are correctly treated as related-but-distinct rather than the same bug. No overstatement stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 6, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36010", + "issue:38175", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43632", + "issue:43638", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43720", + "issue:43824", + "issue:43828", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44530", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44683", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:45030", + "issue:45325", + "issue:45584", + "issue:45663", + "issue:45672", + "issue:45685", + "issue:45812", + "issue:45823", + "issue:45834", + "issue:45902", + "issue:46402", + "issue:46489", + "issue:46515", + "issue:46858", + "issue:46899" + ], + "soft_pairs": [ + "issue:24643|issue:30064", + "issue:44530|issue:44863", + "issue:44863|issue:46402", + "issue:43366|issue:43828", + "issue:45325|issue:45902", + "issue:43596|issue:46515", + "issue:43366|issue:43531", + "issue:44155|issue:45584", + "issue:45663|issue:45823", + "issue:43638|issue:46515", + "issue:43950|issue:45834", + "issue:43366|issue:44589", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:44910|issue:46858", + "issue:43716|issue:44560", + "issue:43645|issue:43646", + "issue:43550|issue:44910", + "issue:44560|issue:45685", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:44683|issue:46515", + "issue:44560|issue:44918", + "issue:45812|issue:46489", + "issue:45823|issue:46515", + "issue:44661|issue:45823", + "issue:38175|issue:43994", + "issue:44560|issue:44805", + "issue:45672|issue:46899", + "issue:43632|issue:45834", + "issue:44877|issue:45030", + "issue:42491|issue:43720", + "issue:43638|issue:44661", + "issue:43596|issue:44661" + ], + "prepared_review_unit_hash": "88f32c22a73e1c1ad918c01b423218d43f98b32f4ae0a25eadfd890d2c651799", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38069, + "estimated_input_tokens": 9518, + "estimated_eval_tokens": 19292 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a mixed cluster of unrelated issues, with a few small duplicate-like subgroups: custom model initialization regressions in Transformers 5.x, Siglip2 wrong outputs, and Gemma4 4-bit multimodal dtype-cast failures.", + "confidence": 0.74, + "canonical_issue_reason": "issue:43646 is the clearest broad statement of the custom model initialization regression, while issue:43645 is just a notebook-specific manifestation.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best representative issue in this set because it is general, concise, and directly captures the shared initialization bug surface.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Different problems: DeepSpeed training weight-shape error vs image processor handling void segmentation maps." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Both are model/runtime failures, but the model families and failure mechanisms differ (Qwen3.5 cache path vs NemotronH checkpoint loading)." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "Different loading bugs: NemotronH checkpoint compatibility vs Sam3 text encoder weights not being loaded." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "Unrelated model-specific issues: GGUF gpt-oss support vs Phi-tiny-MoE autocast dtype mismatch." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Both involve Qwen variants and RoPE/state handling, but they target different code paths and symptoms." + }, + { + "left": "issue:43596", + "right": "issue:46515", + "accept": false, + "reason": "Different classes of failure: DeepSpeed ZeRO3/Bert initialization vs a torch version incompatibility report." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Same broad model family, but one is GGUF architecture support and the other is sliding-window behavior; not the same bug." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Different audio-generation bugs: track-to-track embedding leak vs empty-transcription handling after special-token alignment." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Both concern Gemma4, but one is an FSDP2 shared_kv_states bug and the other is a device_map forward mismatch." + }, + { + "left": "issue:43638", + "right": "issue:46515", + "accept": false, + "reason": "Specific DeepSpeed ZeRO3/Bert index error is not the same as a generic torch version incompatibility report." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer corruption is unrelated to Kosmos2.5 long-OCR index errors." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF gpt-oss support and float8 storage lookup errors are different loading/runtime issues." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Importing a missing class and meta-device model loading failures are different code paths." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 failure and meta-device loading failure are different bugs, even if both are Transformers v5 regressions." + }, + { + "left": "issue:44910", + "right": "issue:46858", + "accept": false, + "reason": "Both touch cache/position handling, but one is Qwen3.5 flash-attention on 3D position_ids and the other is static-cache seq-length arithmetic on non-CUDA backends." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Different models and symptoms: Mistral-3 image preprocessor dtype mismatch vs Qwen3-vl-embedding video StopIteration." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same underlying custom-model initialization regression; the notebook-specific report is a narrower manifestation of the broader issue." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different attention/compile failures on different models and kernels; not the same concrete bug." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Unrelated failures: Qwen3-vl video input iteration vs MPS histogram on Int in MoE code." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "Different subsystems and failures; sharing a theme of model support is too broad to merge." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "Importing GenerationMixin and Qwen3Moe FP8 failures are unrelated bugs." + }, + { + "left": "issue:44683", + "right": "issue:46515", + "accept": false, + "reason": "A torch>=2.9 flex_attention compilation regression is not the same as a torch 2.4.1 incompatibility report." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Both involve Qwen/VL embeddings, but one is video input StopIteration and the other is TRL SFT unpacking failure." + }, + { + "left": "issue:45812", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and likely different mapping tables/code paths." + }, + { + "left": "issue:45823", + "right": "issue:46515", + "accept": false, + "reason": "Gemma4 device-mapping bug and torch 2.4.1 incompatibility are unrelated." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer-mapping validation failure is unrelated to Gemma4 PLE device mismatch." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same model and same wrong-output symptom for siglip2-base-patch16-224; these look like duplicate reports of the same bug." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "StopIteration in Qwen3-vl video handling is unrelated to a tensor-mask shape mismatch error." + }, + { + "left": "issue:45672", + "right": "issue:46899", + "accept": true, + "reason": "Both point to the same Gemma4 4-bit multimodal dtype-casting problem affecting audio/vision embedders." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "Custom-model init flag regression and Kosmos2.5 long-OCR index errors are different bugs." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both are config-validation regressions, but they are on different model configs and do not look like one shared bug." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "LoRA/Qwen3_moe loading on hf4.x and BitNet packed-weight loading are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed/Bert initialization index error and tokenizer-mapping-name validation are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "ZeRO3/Bert init failure and tokenizer mapping validation are unrelated bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it identifies the broad custom-model initialization cluster while treating the other pairs as only similar or unrelated. The accepted duplicate-like pairs are well supported by the titles and the reasons avoid overstating evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 7, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44075", + "issue:44164", + "issue:44291", + "issue:44387", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44910", + "issue:44960", + "issue:44977", + "issue:44991", + "issue:45276", + "issue:45292", + "issue:45313", + "issue:45335", + "issue:45357", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45507", + "issue:45588", + "issue:45812", + "issue:46050", + "issue:46459", + "issue:46468", + "issue:46512", + "issue:46846", + "issue:46897", + "issue:46901" + ], + "soft_pairs": [ + "issue:43317|issue:44756", + "issue:44661|issue:44805", + "issue:38175|issue:39692", + "issue:43644|issue:44534", + "issue:41720|issue:42915", + "issue:42491|issue:45276", + "issue:42915|issue:44756", + "issue:44960|issue:44977", + "issue:44451|issue:44488", + "issue:44863|issue:46846", + "issue:46459|issue:46512", + "issue:45313|issue:45357", + "issue:44488|issue:44991", + "issue:45507|issue:46468", + "issue:43644|issue:43645", + "issue:43257|issue:43854", + "issue:39692|issue:43994", + "issue:43856|issue:43873", + "issue:42491|issue:45292", + "issue:43856|issue:44387", + "issue:43854|issue:43931", + "issue:43708|issue:44075", + "issue:42491|issue:45335", + "issue:44960|issue:46050", + "issue:46897|issue:46901", + "issue:45488|issue:45812", + "issue:45362|issue:45406", + "issue:44164|issue:44291", + "issue:45276|issue:45335", + "issue:44910|issue:45588" + ], + "prepared_review_unit_hash": "19124452ae96b7e598086f86b40a037ed7a6784700f152e914f6586b1fe1440f", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36655, + "estimated_input_tokens": 9164, + "estimated_eval_tokens": 18584 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated model-specific regressions and loading/quantization issues. The only clear duplicate pair is the Transformers v5 non-persistent-buffer junk bug; the rest share broad themes like model loading, tokenizer behavior, or quantization but do not look like the same underlying defect.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43644 is the best canonical issue because it states the core regression most clearly and broadly; 44534 is essentially the same bug with a slightly different title.", + "canonical_pr_reason": null, + "best_issue_reason": "43644 is the strongest representative of the duplicate subcluster and the most general statement of the shared bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Different problems: gpu+cpu offload/dequantized loading vs Strix Halo mmap/OOM." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Unrelated failure modes: tokenizer mapping-name registration vs mask/tensor shape mismatch." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Different SigLIP2 issues: wrong probabilities vs documentation example/model-processor errors." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying Transformers v5 regression about non-persistent buffers being filled with junk; title wording differs only slightly." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both are Qwen3-family, but one is auto device mapping CUDA assert and the other is FineGrainedFP8Config failure; not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Different code paths and symptoms: LoRA checkpoint compatibility across branches vs resize_token_embeddings not updating Gemma4 embeddings." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Different model/setting: Qwen3Moe FP8 failure vs mmap-related OOM on Strix Halo." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "Different models and triggers: GLM5 issue vs Qwen3.5 flash-attention generation bug." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading regressions, but they concern different repositories and no shared concrete code-path is evident." + }, + { + "left": "issue:44863", + "right": "issue:46846", + "accept": false, + "reason": "Different bugs: NemotronH checkpoint loading vs Qwen3.5 tensor-parallel plan missing linear_attn entries." + }, + { + "left": "issue:46459", + "right": "issue:46512", + "accept": false, + "reason": "Both involve dtype handling, but one is loading composite checkpoints and the other is from_config stale config metadata." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Different issues: ZeRO-3 weight loading for language_model vs incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/model loading regressions, but they affect different model families and likely different root causes." + }, + { + "left": "issue:45507", + "right": "issue:46468", + "accept": false, + "reason": "Different models and failures: invalid method call in GraniteMoEHybrid vs broken Nemotron example." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "One is a buffer-initialization regression; the other is a Jupyter custom-model initialization breakage." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Different symptoms: Qwen3 MoE weights conversion with accelerate+deepspeed vs GLM-4.7-Flash unit-test loading." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Similar model family, but one is docs example errors and the other is nonsensical outputs with AutoModel/pipeline." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both mention memory/quantization, but one is Qwen3 MoE training memory use and the other is offloading behavior with quantization." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "Different concerns: checkpoint compatibility across branches vs resize_token_embeddings not propagating to output embeddings." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both are memory-related, but one is training memory usage and the other is increased reserved CUDA memory under int4 quantization." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different loading failures for different models; no clear shared code-path." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer checkpoint-resume step math vs optimizer SGD args being ignored are unrelated." + }, + { + "left": "issue:42491", + "right": "issue:45335", + "accept": false, + "reason": "Both mention embedding/token resizing, but they target different model internals and do not read like the same bug." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "Different issues: GLM5 model behavior vs sinq quantized save/load breakage." + }, + { + "left": "issue:46897", + "right": "issue:46901", + "accept": false, + "reason": "Same high-level pattern (double-shift training loss), but different model implementations and no evidence one fix would cover both." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and different failure mechanisms." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different runtime failures: Qwen3.5 chat crash vs Gemma4Processor missing _tokenizer in transformers serve." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Both involve save/load behavior, but extra_state handling is distinct from the _is_hf_initialized init_empty_weights TypeError." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": false, + "reason": "Related resize_token_embeddings symptom, but different model families and embedding destinations; not enough evidence of one shared bug." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both are flash-attention related, but one is a packed-sequence misinterpretation and the other is an s_aux=None AttributeError." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the only accepted duplicate pair (43644/44534) is strongly supported by near-identical titles. The remaining soft-edge rejections are conservative and do not overclaim shared root causes." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 8, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42796", + "issue:42947", + "issue:43335", + "issue:43381", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43854", + "issue:43873", + "issue:43957", + "issue:44079", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44488", + "issue:44756", + "issue:44863", + "issue:44877", + "issue:44912", + "issue:44928", + "issue:44960", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45859", + "issue:46326", + "issue:46468", + "issue:46829", + "issue:46894", + "issue:46901" + ], + "soft_pairs": [ + "issue:44402|issue:44488", + "issue:39692|issue:44960", + "issue:41720|issue:44155", + "issue:45663|issue:45759", + "issue:43404|issue:43454", + "issue:42947|issue:45704", + "issue:43425|issue:43957", + "issue:44291|issue:44756", + "issue:43873|issue:44387", + "issue:44079|issue:46829", + "issue:36010|issue:39404", + "issue:44292|issue:44912", + "issue:45663|issue:45859", + "issue:44387|issue:44756", + "issue:44155|issue:45823", + "issue:42947|issue:43381", + "issue:43335|issue:45663", + "issue:43957|issue:44756", + "issue:44863|issue:46468", + "issue:45276|issue:45292", + "issue:43381|issue:44928", + "issue:44479|issue:44877", + "issue:44164|issue:44756", + "issue:45684|issue:45698", + "issue:42796|issue:46326", + "issue:45588|issue:45663", + "issue:43854|issue:45216", + "issue:45292|issue:45335", + "issue:44402|issue:44451", + "issue:46894|issue:46901" + ], + "prepared_review_unit_hash": "dc4143c817a652e1916b30526e7c36826335c6cdaf2ff8654740b78bb0f37d52", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36138, + "estimated_input_tokens": 9035, + "estimated_eval_tokens": 18326 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous issue reports with one clear duplicate pair around `resize_token_embeddings` behavior; the rest are similar-by-theme at best but not the same concrete bug.", + "confidence": 0.72, + "canonical_issue_reason": "`issue:45292` is the cleanest representative of the only strong duplicate thread here: it states the generic `resize_token_embeddings` / `output_embeddings` failure, which is the broader form of the more model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:45292` is the best global issue choice because it is the most general and concise formulation of the shared embedding-resize bug, making it the best canonical anchor for the accepted duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading complaints, but they describe different checkpoints and failure modes, not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Unrelated models and problems: SigLIP2 docs vs GLM5 loading." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Different subsystems and symptoms: Qwen auto device mapping crash vs AudioFlamingo3 batched inference leakage." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Both touch model internals, but one is an FSDP attention-state bug and the other is `AutoModelForCausalLM.from_config` unwrapping; not the same issue." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both mention `lm_head` tying, but they affect different model families and do not clearly share one concrete code-path bug." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/LoRA is unrelated to the T5 apex RMSNorm memory leak." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version compatibility and `meta`-device loading are different problems." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "`init_empty_weights` argument handling and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both involve quantization, but one is offloading behavior and the other is int4 memory regression causing OOM." + }, + { + "left": "issue:44079", + "right": "issue:46829", + "accept": false, + "reason": "Different output/container bugs in different components." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Import path regression vs Whisper pipeline `return_language` behavior are not the same bug." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Both are quantization-related, but they concern different formats and failure modes." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Different model-specific failures: FSDP attention state vs CPU offload incompatibility." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Both mention OOM, but one is int4 CUDA reserved memory and the other is mmap on Strix Halo." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "Different bug classes: embedding/token leakage vs device mismatch in Gemma4 PLE." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective under PEFT LoRA and the other is a mode restriction in eval; not the same bug." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "Different model/config bugs with no shared concrete failure path." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading regression is not the same as mmap/OOM behavior." + }, + { + "left": "issue:44863", + "right": "issue:46468", + "accept": false, + "reason": "Same model family, but one is checkpoint loading and the other is an example script; too little evidence they are the same underlying fix." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "`issue:45276` explicitly includes the `output_embeddings` failure covered by `issue:45292`; this is the same resize-token-embedding bug, with 45276 just being a broader/more specific report." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Both mention gradient-related instability, but one is checkpointing in eval mode and the other is RLHF NaNs from Qwen3.5 position IDs." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Different families of regressions: video-input handling vs strict config loading." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "Save/load `extra_state` handling is unrelated to mmap-based OOM." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load of custom modules, but one is permissions propagation and the other is loading the wrong custom module." + }, + { + "left": "issue:42796", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer BPE-marker handling and `continue_final_message` are related only at a broad tokenizer level, not the same concrete bug." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sinkless-model crash and Gemma4 FSDP state rebuilding are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different loading/regression reports for different models and APIs." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": false, + "reason": "Both are about `resize_token_embeddings`, but they target different model implementations (`output_embeddings` vs `decoder.embed_tokens`) and are not clearly the same concrete bug." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Different model-loading reports with different root causes." + }, + { + "left": "issue:46894", + "right": "issue:46901", + "accept": false, + "reason": "Both describe a double-shift training-loss pattern, but they are model-specific reports without enough evidence of one shared concrete code-path fix." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft edge, issue:45276 vs issue:45292, is supported by the titles and strong similarity, and the rest of the verdicts avoid overclaiming shared bugs. The canonical choice for 45292 is defensible as a concise representative of the resize_token_embeddings/output_embeddings thread." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 9, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:43299", + "issue:43425", + "issue:43452", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44479", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44671", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44898", + "issue:44912", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45479", + "issue:45701", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:46050", + "issue:46315", + "issue:46326", + "issue:46491", + "issue:46665", + "issue:46821", + "issue:46869", + "issue:46894", + "issue:46897", + "issue:46899" + ], + "soft_pairs": [ + "issue:43646|issue:43708", + "issue:43645|issue:44164", + "issue:43828|issue:43957", + "issue:44589|issue:44756", + "issue:46315|issue:46491", + "issue:45812|issue:46315", + "issue:43425|issue:43828", + "issue:39404|issue:45701", + "issue:44530|issue:45084", + "issue:39404|issue:46665", + "issue:43611|issue:43645", + "issue:43425|issue:46869", + "issue:44683|issue:45823", + "issue:43425|issue:45823", + "issue:43425|issue:43716", + "issue:44479|issue:44805", + "issue:43906|issue:44530", + "issue:44589|issue:45084", + "issue:43901|issue:44509", + "issue:43994|issue:44079", + "issue:36032|issue:43452", + "issue:44265|issue:44898", + "issue:39404|issue:44912", + "issue:46050|issue:46326", + "issue:44164|issue:45759", + "issue:43906|issue:46326", + "issue:44671|issue:44991", + "issue:44991|issue:45081", + "issue:45823|issue:46899", + "issue:43299|issue:43931", + "issue:45479|issue:46894", + "issue:43856|issue:44756", + "issue:45479|issue:46897", + "issue:44164|issue:46821" + ], + "prepared_review_unit_hash": "475bb83dc03a63ccc0569b9ed97992ab757c01a11a517005917f9cd3900874d4", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38136, + "estimated_input_tokens": 9534, + "estimated_eval_tokens": 19324 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items form a loose transformer-v5 regression cluster, but the soft pairs mostly describe distinct bugs, model families, or code paths rather than true duplicates. No PRs were provided.", + "confidence": 0.72, + "canonical_issue_reason": "issue:43611 is the best representative because it describes a concrete, broadly relevant v5 loading regression and is more general than the highly model-specific reports around it.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is the clearest single issue to anchor the cluster; it is specific enough to be actionable yet broad enough to stand in for related v5 model-loading breakages.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Different problems: custom-model initialization vs trainer checkpoint resumption/max_steps calculation." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook custom-model initialization and extra_state save/load handling are separate failure modes." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "One is an autocast dtype mismatch for a specific model; the other is meta-device loading breakage." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup and mmap OOM on Strix Halo are unrelated runtime issues." + }, + { + "left": "issue:46315", + "right": "issue:46491", + "accept": false, + "reason": "Tokenizer conversion slowdown is not the same as CodeLlama space-preservation regression." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Wrong token IDs for Granite is a correctness regression, not the convert_tokens_to_ids performance bug." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility is too broad and does not match the specific autocast mismatch bug." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline return_language behavior and generic tokenization changes are different regressions." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3.5 PagedAttentionCache crash and template-node compilation failure are different code paths." + }, + { + "left": "issue:39404", + "right": "issue:46665", + "accept": false, + "reason": "Pipeline return_language regression is unrelated to from_pretrained behavior inside Python classes." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both involve loading/custom models, but the concrete bugs and triggers are different." + }, + { + "left": "issue:43425", + "right": "issue:46869", + "accept": false, + "reason": "A torch-version incompatibility is not the same as YOLO multi-GPU device-map failure." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on new torch versions and Gemma4 device placement are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Generic torch incompatibility does not identify the same device-map bug." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch compatibility and image preprocessor/model dtype mismatch are distinct." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video-input regression for Qwen models and mask-shape IndexError are different symptoms and likely different fixes." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "Isolated reproduction of one prior issue is not the same bug as the PagedAttentionCache linear_attention crash." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage-object errors are unrelated to template compilation errors." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Docs mentioning deprecated pipeline tasks is separate from behavior/docs mismatch for return_all_scores." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Siglip2 wrong outputs and ModelOutput key assignment are different underlying problems." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5 tokenizer special-tokens conflict is unrelated to gguf_file loading failures." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export exportability with torch_compilable_check and Perceiver interpolation failure are not the same bug." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper return_language regression and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized save/load breakage and mistral_common continue_final_message are different features and code paths." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "extra_state save/load handling is unrelated to AutoModelForCausalLM config unwrapping." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "A reproduction issue for an older bug does not match the mistral_common tokenizer behavior change." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked-LM outputs and tokenizer-loading failure for EMBEDDIA/est-roberta are separate." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer-related, but one is a loading failure and the other is a regex patch crash for Mistral." + }, + { + "left": "issue:45823", + "right": "issue:46899", + "accept": false, + "reason": "Gemma4 device-map mismatch and Gemma4 multimodal 4-bit dtype casting are different failures." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both concern Qwen3-VL loading, but the reported failures are different enough that they should not be merged." + }, + { + "left": "issue:45479", + "right": "issue:46894", + "accept": false, + "reason": "Zero-loss classification bug and training-loss double-shift bug are different model/task failures." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE training memory usage and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:45479", + "right": "issue:46897", + "accept": false, + "reason": "Single-label classification zero-loss regression is not the same as Florence2 double-shift training bug." + }, + { + "left": "issue:44164", + "right": "issue:46821", + "accept": false, + "reason": "extra_state persistence and TimesFM window_size AttributeError are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 10, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:40990", + "issue:42175", + "issue:42796", + "issue:42915", + "issue:43317", + "issue:43404", + "issue:43421", + "issue:43452", + "issue:43482", + "issue:43541", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43746", + "issue:43883", + "issue:44206", + "issue:44220", + "issue:44451", + "issue:44462", + "issue:44488", + "issue:44530", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44811", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45507", + "issue:45584", + "issue:45859", + "issue:46178", + "issue:46285", + "issue:46326", + "issue:46402", + "issue:46489" + ], + "soft_pairs": [ + "issue:43596|issue:44683", + "issue:44488|issue:44987", + "issue:44488|issue:45081", + "issue:43611|issue:43644", + "issue:44462|issue:46489", + "issue:42796|issue:43404", + "issue:36032|issue:42915", + "issue:46285|issue:46402", + "issue:39404|issue:44671", + "issue:43716|issue:45084", + "issue:43643|issue:44530", + "issue:44737|issue:45859", + "issue:44661|issue:44683", + "issue:40990|issue:43421", + "issue:42915|issue:44530", + "issue:44451|issue:44987", + "issue:43644|issue:43646", + "issue:34567|issue:40444", + "issue:43541|issue:45084", + "issue:40444|issue:46326", + "issue:42175|issue:46178", + "issue:43582|issue:43632", + "issue:44206|issue:44220", + "issue:44811|issue:45584", + "issue:36032|issue:43645", + "issue:43452|issue:43482", + "issue:43883|issue:44488", + "issue:44987|issue:44991", + "issue:43317|issue:43482", + "issue:45084|issue:45507", + "issue:43746|issue:46326" + ], + "prepared_review_unit_hash": "b5794be8d42c578a6bef2b0b4184f3c3c73aa8a9d3da29c8c55cdbfa6305787e", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37017, + "estimated_input_tokens": 9255, + "estimated_eval_tokens": 18766 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose grab-bag of unrelated Transformers regressions and model/tokenizer edge cases, with only weak thematic overlap around v5 loading behavior. None of the soft pairs are strong enough duplicates to merge.", + "confidence": 0.91, + "canonical_issue_reason": "No single issue is a safe canonical for the cluster: the items span distinct bugs across tokenizers, model initialization/loading, audio, quantization, and runtime/device behavior.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the broadest representative of the cluster\u2019s recurring v5 regression theme, but it is still too generic to absorb the others as duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Different failures and code paths: Deepspeed/ZERO3 BertModel indexing vs torch>=2.9 flex_attention compilation." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading regressions, but they concern different models and likely different tokenizer/model paths." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "BERT model loading failure is not the same bug as the Mistral tokenizer regex patch crash." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffer initialization corruption." + }, + { + "left": "issue:44462", + "right": "issue:46489", + "accept": false, + "reason": "Both involve tokenizer loading/output, but the titles point to different model-specific bugs and different symptoms." + }, + { + "left": "issue:42796", + "right": "issue:43404", + "accept": false, + "reason": "Tokenizer BPE-marker handling and lm_head weight tying are separate Mistral3 issues." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "T5 tokenizer load error and Qwen3Moe FineGrainedFP8 failure are unrelated." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "CLIP structure inconsistency is unrelated to Sam3 text encoder weights not loading." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language regression and CamemBERT masked-LM prediction regression are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Image preprocessor/model dtype mismatch is a different bug from the template-node compile error." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "trust_remote_code config field loss and Qwen3.5 paged-attention invalid group type are unrelated." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "Both mention device placement, but they affect different models and different code paths." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer mapping-name validation and flex_attention compilation are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity regression on a benchmark and runtime special-token post-processor updates are not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Same broad Qwen family, but FP8 config failure and paged-attention cache crash are different issues." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are latest-version loading failures, but for different repositories and likely different root causes." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Related to v5 initialization, but the symptoms differ: buffer corruption vs custom model initialization breakage." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "TrainerState token accounting and Qwen2.5-VL iterable dataset failures are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing failure is unrelated to the non-template-node compile error." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Multimodal training with multiple images and Whisper tokenizer continue_final_message are different bugs." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Backend packaging/installation issue and tensorflow_text backend mapping error are related to dependencies, not the same defect." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "AppleSilicon allocator warmup TypeError and _is_hf_initialized regressions are separate initialization problems." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Both are audio-feature related, but the titles do not establish the same root cause with enough confidence." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Whisper batch_decode skip_special_tokens behavior and empty-transcription generation after align_special_tokens are different symptoms." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer class-method conflict and custom model initialization in notebooks are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "GGUF from_pretrained breakage and Qwen2.5-GGUF loading failure are both GGUF-related but not clearly the same bug." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Molmo tied-weights attribute error and BERT model loading failure are unrelated." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are transformer v5 tokenizer-load failures, but they affect different models and likely different tokenizer issues." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Dequantized model offload failure is unrelated to Qwen2.5-GGUF loading failure." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation failure and GraniteMoEHybrid invalid method calls are different bugs." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT local-checkpoint loading for GraniteSpeech is unrelated to Whisper continue_final_message behavior." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 11, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28530", + "issue:34567", + "issue:36032", + "issue:40444", + "issue:41720", + "issue:42915", + "issue:43089", + "issue:43262", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43531", + "issue:43540", + "issue:43632", + "issue:43645", + "issue:43646", + "issue:43742", + "issue:43792", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44849", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45987", + "issue:46315", + "issue:46661", + "issue:46821", + "issue:46869", + "issue:46899" + ], + "soft_pairs": [ + "issue:43262|issue:43377", + "issue:43646|issue:44534", + "issue:43645|issue:46661", + "issue:44403|issue:45081", + "issue:44849|issue:45325", + "issue:42915|issue:45084", + "issue:43377|issue:43819", + "issue:43883|issue:44991", + "issue:43792|issue:46869", + "issue:43404|issue:44466", + "issue:44464|issue:45987", + "issue:43994|issue:44610", + "issue:44987|issue:45081", + "issue:28530|issue:34567", + "issue:44008|issue:46899", + "issue:42915|issue:44291", + "issue:44451|issue:45081", + "issue:36032|issue:43742", + "issue:44062|issue:44987", + "issue:44265|issue:44610", + "issue:43089|issue:45987", + "issue:43322|issue:43540", + "issue:44479|issue:45325", + "issue:42915|issue:44589", + "issue:36032|issue:41720", + "issue:43632|issue:46821", + "issue:44991|issue:46315", + "issue:43531|issue:44877", + "issue:45216|issue:45357", + "issue:40444|issue:45325" + ], + "prepared_review_unit_hash": "d226bb892b17e40552473eca605b844ec34037739f33a371a40030df4932cf9e", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36937, + "estimated_input_tokens": 9235, + "estimated_eval_tokens": 18726 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a loose topical cluster of unrelated Transformer regressions. I\u2019d keep only the Qwen3.5 save_pretrained pair (45216\u219445357) as a clear duplicate; the tokenizer AddedToken and Qwen VL rope-index pairs look too weakly supported to merge confidently.", + "confidence": 0.77, + "canonical_issue_reason": "Issue 45216 is the clearest canonical representative: it states the Qwen3.5 save_pretrained regression directly, and 45357 reads like a narrower follow-up report of the same incorrect visual-encoder-keys problem.", + "canonical_pr_reason": null, + "best_issue_reason": "45216 is the best standalone issue anchor because it describes a concrete regression with a direct duplicate/follow-up report, making the bug scope easiest to interpret.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Both are audio-related, but they describe different failures: chat-template sampling-rate defaulting vs batched padding-mask behavior in MIMI." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization breakage and non-persistent buffer corruption are different code paths and different bugs." + }, + { + "left": "issue:43645", + "right": "issue:46661", + "accept": false, + "reason": "Both involve model loading, but the failure modes differ: notebook custom-model initialization vs missing config_class on remote-code models." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Unnecessary load-time noise and a Mistral tokenizer regex patch crash are unrelated symptoms with different fixes." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "These are different Qwen regressions: output_hidden_states handling in Qwen3.5 versus rope-index temporal-position scaling in Qwen2.5-VL." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "FineGrainedFP8Config failure and non-template-node compilation are distinct issues." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "Both are multimodal/audio-model bugs, but one is missing padding-mask support and the other is DAC from_latents disagreeing with forward due to missing STE." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Missing all_tied_weights_keys in Molmo is unrelated to the tokenizer-loading regression for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43792", + "right": "issue:46869", + "accept": false, + "reason": "Whisper load failure and YOLO multi-GPU auto device-map failure are separate model-specific problems." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "They both mention tied weights, but one is a model-specific tying bug and the other is a serialization inconsistency across device placement." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is not the same as StaticCache.get_seq_length returning the wrong type." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Siglip2 AutoModel/pipeline nonsense outputs and OmDet-Turbo processor size mismatch are different vision bugs." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "A model-specific tokenizer load regression and a Mistral tokenizer regex patch crash are not the same underlying bug." + }, + { + "left": "issue:28530", + "right": "issue:34567", + "accept": false, + "reason": "Early stopping metric selection and num_input_tokens_seen not updating are unrelated Trainer issues." + }, + { + "left": "issue:44008", + "right": "issue:46899", + "accept": false, + "reason": "Gemma3n variable collision in forward and Gemma4 multimodal dtype-cast corruption are different bugs." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "FineGrainedFP8Config failure and init_empty_weights/_is_hf_initialized argument breakage are separate initialization problems." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "ScandiBERT loading failure is not the same as the Mistral regex tokenizer crash." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5 add_special_tokens method conflict and MobileLLM key error are different tokenizer/model-loading issues." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "The evidence is too thin: one side is a generic tokenizer-loading regression and the other is a specific AddedToken duplicate-special TypeError." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export / torch_compilable_check failure is unrelated to the OmDet-Turbo processor size regression." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation overhead from GPU syncs/dispatch is a performance issue, not the seq_length return-type bug." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava-Next loading segfault and Qwen3OmniMoe video-input ValueError are unrelated multimodal failures." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "This mixes a broad video-input regression with a still-image rope-index bug; the shared Qwen VL theme is not enough to treat them as the same underlying issue." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "FineGrainedFP8Config failure and missing Float8 storage object are different quantization/runtime errors." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer method-name collision and Qwen3 auto device-map cuda assert are unrelated." + }, + { + "left": "issue:43632", + "right": "issue:46821", + "accept": false, + "reason": "_is_hf_initialized breakage and TimesFM window_size AttributeError are distinct v5 regressions." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer-related, but one is a load failure and the other is a performance regression in convert_tokens_to_ids." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and granite_speech strict-config loading are unrelated issues." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "These are the same Qwen3.5 save_pretrained regression; the later report just narrows the symptom to incorrect visual encoder keys." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Multiple-images-per-prompt finetuning failure and still-image temporal-position scaling are different Qwen2.5-VL bugs." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The overall framing is conservative and grounded, but two of the accepted soft-edge matches are too assertive for the evidence in the packet. In particular, 44062\u219444991 is only supported by a generic tokenizer-loading regression on one side and a specific AddedToken TypeError on the other, and 44479\u219445325 mixes a broad video-input regression with a still-image rope-index bug. Those should stay tentative unless the issue bodies clearly line up. The 45216\u219445357 duplicate looks fine, but because the cluster summary relies on the other two accepted pairs being duplicates, I\u2019d reject for overstating evidence." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 12, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:40444", + "issue:40990", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43381", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43716", + "issue:43746", + "issue:43792", + "issue:43856", + "issue:44079", + "issue:44220", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44509", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44991", + "issue:45081", + "issue:45491", + "issue:45538", + "issue:45542", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:46491", + "issue:46710", + "issue:46899", + "issue:46909" + ], + "soft_pairs": [ + "issue:43611|issue:44534", + "issue:43425|issue:44403", + "issue:43716|issue:46899", + "issue:43296|issue:43531", + "issue:45812|issue:46491", + "issue:44220|issue:44683", + "issue:42796|issue:43746", + "issue:43381|issue:45923", + "issue:44403|issue:44991", + "issue:43632|issue:44756", + "issue:33666|issue:40444", + "issue:40990|issue:44811", + "issue:43792|issue:44589", + "issue:44291|issue:44589", + "issue:42796|issue:46710", + "issue:44079|issue:45663", + "issue:35707|issue:40444", + "issue:43482|issue:43525", + "issue:43421|issue:44568", + "issue:43632|issue:44534", + "issue:43482|issue:45538", + "issue:43531|issue:44291", + "issue:41553|issue:43582", + "issue:42175|issue:45542", + "issue:41720|issue:45081", + "issue:43317|issue:43856", + "issue:44079|issue:44534", + "issue:44462|issue:45812", + "issue:43377|issue:45491", + "issue:44509|issue:46909" + ], + "prepared_review_unit_hash": "3fb937f89ee49f87516083b647d53353285cf2c49de871aae59ca596513d4bb7", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36621, + "estimated_input_tokens": 9156, + "estimated_eval_tokens": 18568 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issues; the soft pairs only share broad themes like v5 regressions, tokenizers, or multimodal training, but not the same underlying bug/change. I would not merge any of the suggested pairs.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out because this is a heterogeneous set; the soft pairs do not point to one shared defect. If forced to pick an anchor, issue:44534 is the most generic core-regression report, but it is not a duplicate hub.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global best duplicate representative here. The cluster is too mixed, and the candidate pairs are only loosely related by subsystem, not by exact bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Both are Transformers v5 regressions, but they describe different failures: base_model_prefix loading vs junk-filled non-persistent buffers." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Unrelated problems: Torch 2.10 compatibility versus noisy logging during model loading." + }, + { + "left": "issue:43716", + "right": "issue:46899", + "accept": false, + "reason": "Both are multimodal dtype issues, but one is image preprocessor/model mismatch and the other is 4-bit unsafe casting in Gemma4 embedders." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different models and failure modes: PaddleOCR-VL load failure vs Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:45812", + "right": "issue:46491", + "accept": false, + "reason": "Both concern tokenization, but one is Granite token IDs and the other is a CodeLlama leading-space round-trip regression." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction bug versus compiled flex_attention failure on newer Torch versions." + }, + { + "left": "issue:42796", + "right": "issue:43746", + "accept": false, + "reason": "Tokenizer BPE-marker handling and GraniteSpeech PEFT checkpoint loading are separate bugs." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Gradient checkpointing in eval mode is not the same as a missing supports_gradient_checkpointing flag on a trust_remote_code model." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Generic loading noise and a specific AutoTokenizer regression for EMBEDDIA/est-roberta are unrelated." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "_is_hf_initialized regression and Strix Halo mmap OOM are different code paths and symptoms." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both involve Qwen2-VL-family training, but one is multi-GPU training and the other is IterableDataset/multiple-images finetuning." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Perplexity regression on gpt-oss-20b is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper model load/runtime failure and missing Float8 storage are not the same bug." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized TypeError is distinct from the Float8 storage-object error." + }, + { + "left": "issue:42796", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer BPE-marker bug and DeepSeek accuracy regression are different issues, despite both being model-specific regressions." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment and Gemma4 shared_kv_states KeyError are unrelated implementation bugs." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to multi-image iterable finetuning failures." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Both are v5 loading failures, but one is Qwen2.5-GGUF compatibility and the other is missing pad_token_id on Llama4Config." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime special-token post-processor updates and mdeberta-v3-base add_special_tokens behavior are related only at a very broad tokenizer level, not the same concrete bug." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Different Transformers v5 regressions: _is_hf_initialized flag handling versus non-persistent buffer initialization." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading failure and CLIPTokenizer model_max_length overflow are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and init_empty_weights argument handling are distinct issues." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Bad AutoTokenizer error messaging for Voxtral is unrelated to the Apple Silicon TypeError in caching_allocator_warmup." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both mention TensorFlow packaging, but one is missing TF in the torch extra and the other is a tensorboard-only undefined backend error." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Qwen3 auto device mapping cudaErrorAssert and Mistral tokenizer regex patch crashes are different problems." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "GPU+CPU offload loading failure and Qwen3 MoE training memory inefficiency are not the same bug." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment and junk-filled non-persistent buffers are different core behaviors." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Ignoring tokenizer.json and Granite token-ID regression are both tokenizer issues but not the same code-path problem." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "MIMI encoder padding-mask mismatch and Gemma3 NaN embeddings from sliding-window attention are separate model bugs." + }, + { + "left": "issue:44509", + "right": "issue:46909", + "accept": false, + "reason": "Removed pipeline docs lingering in v5 and a broken DIA docs example are distinct documentation issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 13, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42796", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43454", + "issue:43482", + "issue:43493", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43818", + "issue:43819", + "issue:43828", + "issue:43873", + "issue:43881", + "issue:43994", + "issue:44079", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44466", + "issue:44479", + "issue:44488", + "issue:44829", + "issue:44877", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45584", + "issue:45701", + "issue:45823", + "issue:45834", + "issue:46402", + "issue:46821" + ], + "soft_pairs": [ + "issue:43611|issue:43646", + "issue:43296|issue:46821", + "issue:43818|issue:46402", + "issue:44829|issue:45479", + "issue:43644|issue:44079", + "issue:42175|issue:43296", + "issue:43818|issue:45357", + "issue:43493|issue:43994", + "issue:44265|issue:45072", + "issue:40444|issue:45381", + "issue:45325|issue:45834", + "issue:45325|issue:45381", + "issue:43454|issue:44466", + "issue:42915|issue:43482", + "issue:43873|issue:45701", + "issue:43701|issue:45325", + "issue:44291|issue:44877", + "issue:42796|issue:45381", + "issue:43606|issue:44265", + "issue:40990|issue:45584", + "issue:43262|issue:43746", + "issue:44479|issue:45823", + "issue:35707|issue:41720", + "issue:34634|issue:35707", + "issue:43828|issue:46821", + "issue:39692|issue:40444", + "issue:44186|issue:44265", + "issue:43454|issue:43550", + "issue:43606|issue:43819", + "issue:43881|issue:44488", + "issue:41720|issue:45823" + ], + "prepared_review_unit_hash": "f0504f34482e2e1c875c6358df39c3bda87d24430426a870c5c8289ce1a84d0a", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36522, + "estimated_input_tokens": 9131, + "estimated_eval_tokens": 18518 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are only loosely related by model family or regression theme, not the same underlying bug. The only likely duplicate-looking pair is the SigLIP2 implementation issue; the rest should stay separate.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43493 is the strongest canonical issue candidate because it describes the underlying SigLIP2 implementation discrepancy directly and broadly, while issue:43994 reads more like a user-facing symptom of the same problem.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43493 is the best overall issue to anchor this set: it is specific enough to identify the root cause, but broad enough to cover the downstream bad outputs reported elsewhere.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are Transformers 5.0 loading regressions, but one is about `base_model_prefix` loading and the other about custom model initialization; different failure modes and likely different fixes." + }, + { + "left": "issue:43296", + "right": "issue:46821", + "accept": false, + "reason": "Unrelated models and symptoms: PaddleOCR-VL/vLLM loading versus TimesFM `window_size` attribute handling." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Different models and root causes: Video-LLaVA temporal attention/weight sharing versus SAM3 text encoder weights not loading." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both mention degenerate training/loss behavior, but the concrete triggers differ (`flash_attention_3` regression vs `num_labels=1` classification loss logic)." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "One is a non-persistent buffer serialization bug; the other is `ModelOutput` key assignment behavior. Not the same code-path." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Packaging/backend dependency issue versus a model loading failure; no shared bug." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Both involve multimodal model weights, but one is Video-LLaVA architecture/weights and the other is Qwen3.5 save_pretrained key serialization." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both are about SigLIP2 behaving incorrectly in HF vs expected behavior, and the second looks like a downstream symptom of the same implementation discrepancy." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "Different export/inference problems: torch.export with `torch_compilable_check` versus bfloat16 dtype mismatches in specific models." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Both involve Qwen multimodal inputs, but one is iterable-dataset multi-image finetuning and the other is video position-id handling; not the same bug." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Different models and failures: Qwen2.5-VL RoPE/position-id scaling versus Kosmos2.5 OCR index errors." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Closely related Qwen2.5-VL position-id bugs, but one is a still-image temporal-position scaling issue and the other is video input vision_position_ids; too different to merge confidently." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both mention tied `lm_head` weights, but one is a model-specific weight-tying failure and the other is device-dependent serialization inconsistency." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Different models and failure domains: FP8 quantization on Qwen3Moe versus GGUF loading on Qwen2.5." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantization/offload behavior versus tokenization changes; no shared underlying bug." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL position-id computation." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "`init_empty_weights` argument handling and strict config loading for granite_speech are different regressions." + }, + { + "left": "issue:42796", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer BPE-marker handling is unrelated to Qwen2.5-VL video position-id computation." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU offload device mismatch in Bark-small versus export failure with `torch_compilable_check`; different subsystems." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Perplexity regression on GPT-OSS is unrelated to Whisper generation on empty transcription." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Qwen video-input regression and Gemma4 device-map mismatch are separate issues." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation with `inputs_embeds`/`past_key_values` is unrelated to Qwen3 auto device-map CUDA asserts." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice preset handling and progressive generation with embeddings are unrelated bugs." + }, + { + "left": "issue:43828", + "right": "issue:46821", + "accept": false, + "reason": "Phi-tiny-MoE autocast dtype mismatch is unrelated to TimesFM `window_size` attribute errors." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 docs/example issues and Qwen2.5-VL multi-image finetuning failures are different problems." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer crashes on NER/batched padding are unrelated to export behavior with `torch_compilable_check`." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "Weight-tying/serialization bug versus Bamba torch.compile SDPA regression; not the same issue." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Both are model inference mismatches, but one is CPU offload device placement and the other is DAC latent/STE forward mismatch." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Separate loading failures for different models." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Qwen3 auto device-map CUDA assert and Gemma4 PLE device mismatch are different model/runtime problems." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: issue:43493 is the clearest SigLIP2 root-cause issue, and the 43493/43994 pair is the only one that reasonably looks like the same bug rather than just a related symptom. The other soft-edge rejections are appropriately conservative and consistent with the titles shown." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 14, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:26413", + "issue:28530", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42796", + "issue:42915", + "issue:42947", + "issue:43065", + "issue:43257", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43450", + "issue:43454", + "issue:43526", + "issue:43572", + "issue:43720", + "issue:43749", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44186", + "issue:44387", + "issue:44466", + "issue:44534", + "issue:44568", + "issue:44610", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45357", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46209", + "issue:46305", + "issue:46315", + "issue:46402", + "issue:46821" + ], + "soft_pairs": [ + "issue:43377|issue:44610", + "issue:43257|issue:45923", + "issue:44568|issue:45920", + "issue:46305|issue:46315", + "issue:42796|issue:46209", + "issue:43454|issue:45072", + "issue:45823|issue:46821", + "issue:40990|issue:42947", + "issue:43377|issue:44898", + "issue:43454|issue:43720", + "issue:42617|issue:42915", + "issue:43749|issue:44534", + "issue:43317|issue:43873", + "issue:43572|issue:45759", + "issue:43322|issue:46050", + "issue:41720|issue:43322", + "issue:36331|issue:42617", + "issue:44387|issue:45704", + "issue:41720|issue:44977", + "issue:36331|issue:42915", + "issue:44466|issue:45357", + "issue:43526|issue:43883", + "issue:40990|issue:45701", + "issue:43450|issue:45701", + "issue:43881|issue:44991", + "issue:43065|issue:46402", + "issue:44877|issue:44991", + "issue:26413|issue:28530", + "issue:41720|issue:42617", + "issue:44186|issue:45072", + "issue:43873|issue:44756" + ], + "prepared_review_unit_hash": "09ca674e0ec8088d4852ed400f1abd74e8d848c106e4d0ba6d72981fbc3ab240", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36986, + "estimated_input_tokens": 9247, + "estimated_eval_tokens": 18750 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are distinct Transformers bugs spanning tokenization, loading/offload, quantization, vision, and training-trainer regressions. None of the provided soft pairs look like true duplicates of the same underlying bug/change.", + "confidence": 0.92, + "canonical_issue_reason": "If a single issue must represent the cluster, issue 43873 is the broadest umbrella, but the overall set is too diverse for a strong canonical issue choice.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the best representative only because it is a broad loading/offload/quantization report, but it is not a good global duplicate anchor for this mixed cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Both are batched-input bugs, but one is about MIMI encoder padding masks and the other about OmDet-Turbo image size mismatches; different code paths and fixes." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Both mention Qwen/MoE-style model handling, but one is about accelerate+deepspeed conversion of MoE weights and the other is a missing gradient-checkpointing flag on a trust_remote_code model." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and failure modes: missing BOS/EOS insertion vs wrong token IDs across several models." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "Both are performance/behavior regressions in windowing/token processing, but one is an off-by-one streaming mask bug and the other is slow convert_tokens_to_ids for slow tokenizers." + }, + { + "left": "issue:42796", + "right": "issue:46209", + "accept": false, + "reason": "One is a tokenizer BPE-marker handling bug for a specific model; the other is an FP8 inference failure on B200 hardware. No shared underlying defect." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "One concerns a processor/model input-shape mismatch; the other is dtype mismatches in bfloat16 inference. Too different to merge." + }, + { + "left": "issue:45823", + "right": "issue:46821", + "accept": false, + "reason": "Both involve model/config loading, but one is a Gemma4 device-mapping forward mismatch and the other is a TimesFM window_size attribute error. Different concrete bugs." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "One reports high perplexity on a model/dataset combination; the other is gradient checkpointing ineffective with LoRA. They do not describe the same bug." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "Both are batch-related behavior issues, but one is missing padding mask support in MIMI and the other is a Perceiver image-resolution failure. Not the same fix." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "One is a processor shape bug; the other is a BitNet packed-weight unpacking problem during accelerate loading. Different mechanisms and code paths." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Both are runtime failures, but one is about running 3d_parallel.py and the other is a Qwen3Moe FineGrainedFP8Config issue. No clear shared underlying bug." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "Both touch loading/runtime behavior, but one is FSDP CPU RAM efficient loading and the other is non-persistent buffers filled with junk in v5. Distinct issues." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both mention offload/quantization, but one is device_map=auto failing to load a dequantized model with cpu offload, while the other is a broader quantization offloading complaint. Too broad to treat as one bug." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Both are config/model setup regressions, but one is a missing pad_token_idx after a 5.0 update and the other is AutoModelForCausalLM.from_config not unwrapping text_config for Qwen multimodal configs." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Segmentation fault loading Llava Next models is unrelated to sinq quantization breakage after save/load." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Both are model loading failures, but one is a cudaErrorAssert on Qwen3 with auto device mapping and the other is a segmentation fault with Llava Next. Different failure signatures and likely causes." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "One is a Trainer API signature change for compute_loss; the other is a 3d_parallel.py execution problem. No shared bug." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Both are memory-related, but one is CUDA reserved-memory growth under int4 quantization and the other is a T5/apex fused RMSNorm leak. Not the same underlying issue." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Both involve Qwen models, but one is a loading/assert failure and the other is flash-attention generation quality/correctness. Different code paths." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "A Trainer compute_loss signature mismatch is unrelated to Qwen3Moe FP8 configuration failures." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both concern serialization of model components in v5, but one is tied lm_head.weight serialization and the other is incorrect visual encoder keys for Qwen3.5. Different artifacts and fixes." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "One is a BeitImageProcessorFast label-reduction bug; the other is a missing all_tied_weights_keys attribute on MolmoForCausalLM. Unrelated." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "One is a perplexity regression on a specific model/dataset; the other is tokenizer output changing across transformers versions. Not the same defect." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batch-shape issues are distinct from tokenizer-version-dependent tokenization changes." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "One is a glm-4v-9b loading failure; the other is tokenizer loading failure for a specific RoBERTa model. Different components and symptoms." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "A dummy Conv2d in Sam3PixelDecoder is unrelated to Sam3Model text encoder weights not loading." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Strict config rejection for granite_speech is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." + }, + { + "left": "issue:26413", + "right": "issue:28530", + "accept": false, + "reason": "Resume-from-checkpoint epoch iterator failure and early-stopping metric lookup are separate Trainer bugs." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "A Qwen3 auto-device-map cuda assert and a 3d_parallel.py run failure are different problems." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "LayoutLMv2Tokenizer NER/padding crashes are unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel inference." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both are memory/offload themed, but one is quantization offloading behavior and the other is disabling mmap on Strix Halo to avoid OOM. Not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 15, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28530", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43295", + "issue:43316", + "issue:43317", + "issue:43322", + "issue:43404", + "issue:43425", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43646", + "issue:43708", + "issue:43720", + "issue:43881", + "issue:44008", + "issue:44186", + "issue:44361", + "issue:44423", + "issue:44464", + "issue:44610", + "issue:44671", + "issue:44734", + "issue:44743", + "issue:44898", + "issue:44913", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45685", + "issue:45701", + "issue:45823", + "issue:46291", + "issue:46612", + "issue:46665", + "issue:46821", + "issue:46858", + "issue:46869" + ], + "soft_pairs": [ + "issue:46612|issue:46858", + "issue:43550|issue:43720", + "issue:43577|issue:43708", + "issue:43295|issue:45701", + "issue:44186|issue:44610", + "issue:45685|issue:46821", + "issue:43550|issue:43606", + "issue:43646|issue:46665", + "issue:43404|issue:45357", + "issue:44610|issue:44898", + "issue:36010|issue:42617", + "issue:42915|issue:43317", + "issue:44734|issue:46858", + "issue:44423|issue:45406", + "issue:44008|issue:45823", + "issue:46821|issue:46869", + "issue:45632|issue:45684", + "issue:41720|issue:45161", + "issue:28530|issue:33666", + "issue:44464|issue:46858", + "issue:44671|issue:44743", + "issue:43531|issue:43572", + "issue:36331|issue:41720", + "issue:42915|issue:46291", + "issue:44186|issue:44361", + "issue:43316|issue:44913", + "issue:42915|issue:43322", + "issue:43425|issue:46821", + "issue:43526|issue:43881" + ], + "prepared_review_unit_hash": "19ff046977a0f742696b51ae9395930d48e8912bd5c43e166bcf7d459f14ea58", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36487, + "estimated_input_tokens": 9122, + "estimated_eval_tokens": 18500 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous and mostly contains loosely related Transformers bug reports rather than true duplicates. The strongest sub-theme is generation/cache/compile regressions, with issue 46858 as the best anchor, but there is no single canonical duplicate across the whole set.", + "confidence": 0.29, + "canonical_issue_reason": "Issue 46858 is the closest hub because it is a concrete generation/cache regression and is semantically closest to multiple other cache/compile failures in the cluster, even though the overall set is not a single duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46858 is the best overall representative: it is specific, technically grounded, and most central to the recurring runtime/generation-cache failures shown by the soft links.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46612", + "right": "issue:46858", + "accept": false, + "reason": "Both involve generation internals, but one is beam-search cache reorder logic for legacy models and the other is a static-cache/_prefill device-tensor arithmetic bug in compile paths; not the same fix." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Different failure modes: torch.compile with SDPA in Bamba versus packed-weight unpacking during accelerate loading for BitNet." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Blip2 dtype propagation on load and Trainer checkpoint-resume step \u0627\u0644\u062d\u0633\u0627\u0628 are unrelated bugs." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Both mention tokenization-related regressions, but one is a processor/tokenizer API break and the other is version-dependent tokenization differences; not the same underlying bug." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer padding/truncation crash in LayoutLMv2 is unrelated to OmDet-Turbo producing the wrong image resolution." + }, + { + "left": "issue:45685", + "right": "issue:46821", + "accept": false, + "reason": "An MPS histogram kernel error in MoE code is not the same as a TimesFM window_size AttributeError." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both are runtime issues, but one is a Bamba SDPA compile failure and the other is a CPU offload device-mismatch bug in bark-small." + }, + { + "left": "issue:43646", + "right": "issue:46665", + "accept": false, + "reason": "Custom model initialization breakage and from_pretrained behavior inside Python classes are related only at a very high level, not the same concrete bug." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Mistral3 lm_head tying and Qwen3.5 save_pretrained key serialization are different model-state problems." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision-model preprocessing issues, but one is a fixed-size processor/model mismatch and the other is positional-encoding interpolation failure." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "An import error for GenerationMixin is unrelated to not being able to run 3d_parallel.py." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "FineGrainedFP8Config failure in Qwen3Moe and dequantized model offload loading failure are different quantization/offload bugs." + }, + { + "left": "issue:44734", + "right": "issue:46858", + "accept": false, + "reason": "Both touch cache continuation, but one is wrong tensor indexing in serve responses and the other is static-cache seq-length device arithmetic; not mergeable as one fix." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both are serve/multimodal processor regressions, but the concrete errors differ: missing .to() on a string versus Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Gemma3n variable-name collision and Gemma4 PLE device mismatch are distinct forward-pass bugs." + }, + { + "left": "issue:46821", + "right": "issue:46869", + "accept": false, + "reason": "TimesFM window_size handling and YOLO multi-GPU device-map failure are unrelated." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collisions for trust_remote_code and read-only permission propagation in save_pretrained are different filesystem/persistence issues." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Qwen3 auto-device assert and GPT-OSS MoE tensor-parallel failure are separate model/runtime problems." + }, + { + "left": "issue:28530", + "right": "issue:33666", + "accept": false, + "reason": "Early-stopping metric selection and Qwen2-VL multi-GPU training are unrelated training issues." + }, + { + "left": "issue:44464", + "right": "issue:46858", + "accept": false, + "reason": "Both involve compiled generation, but chunked-generation output inconsistency is a different bug than cache sequence-length slice arithmetic on non-CUDA backends." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM prediction drift and Qwen3.5 recurrent-state reset with cache are different model-specific behaviors." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "A sliding-window issue in Qwen3-MoE is unrelated to StableLmConfig missing pad_token_idx after the 5.0 update." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer compute_loss signature mismatch and Qwen3 device-map CUDA assert are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:46291", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and kernels package version/revision import requirement are different dependency/runtime issues." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer failures, but they affect different tokenizers, tasks, and code paths." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and GPTNeoXConfig rotary_pct reload behavior are distinct configuration bugs." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FP8 loading failure and Llava Next segmentation fault are not the same underlying problem." + }, + { + "left": "issue:43425", + "right": "issue:46821", + "accept": false, + "reason": "Torch 2.10 incompatibility and TimesFM window_size AttributeError are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and glm-4v-9b loading failure are different issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 16, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28530", + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:41720", + "issue:42490", + "issue:43329", + "issue:43388", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43644", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:44062", + "issue:44077", + "issue:44387", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44534", + "issue:44589", + "issue:44625", + "issue:44898", + "issue:44991", + "issue:45059", + "issue:45072", + "issue:45357", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:46402", + "issue:46665", + "issue:46821", + "issue:46858" + ], + "soft_pairs": [ + "issue:28530|issue:30064", + "issue:33357|issue:35707", + "issue:29159|issue:36032", + "issue:44387|issue:45701", + "issue:43540|issue:43782", + "issue:43873|issue:45704", + "issue:30064|issue:34634", + "issue:43550|issue:43653", + "issue:43606|issue:43720", + "issue:44534|issue:46665", + "issue:41720|issue:43746", + "issue:43531|issue:45759", + "issue:44492|issue:44509", + "issue:45406|issue:45464", + "issue:42490|issue:43329", + "issue:43388|issue:45059", + "issue:43606|issue:43653", + "issue:34567|issue:35707", + "issue:34634|issue:41720", + "issue:43531|issue:46821", + "issue:43540|issue:46858", + "issue:45759|issue:45859", + "issue:45357|issue:46402", + "issue:44451|issue:44991", + "issue:44077|issue:44625", + "issue:43531|issue:44589", + "issue:43644|issue:46665", + "issue:44062|issue:44451", + "issue:44898|issue:45072", + "issue:44625|issue:45479" + ], + "prepared_review_unit_hash": "f5ad6ed01f44ee91d298cea32671872e27242f869a5a1244c056914e1bfa1306", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36305, + "estimated_input_tokens": 9077, + "estimated_eval_tokens": 18410 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a false-positive mix of unrelated issues across tokenization, loading, vision/multimodal processing, generation, offloading, and docs. None of the soft edges look like true duplicates, and there are no PRs to compare.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the cluster spans multiple unrelated root causes, so there is no single representative bug to merge everything into.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a good global canonical; if forced, issue:43606 is the broadest theme (offloading/quantization), but it still does not represent the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:28530", + "right": "issue:30064", + "accept": false, + "reason": "Early-stopping metric selection is unrelated to image processor handling of void segmentation maps." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are different code paths and failures." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency is not the same as a T5Tokenizer load-time add_special_tokens conflict." + }, + { + "left": "issue:44387", + "right": "issue:45701", + "accept": false, + "reason": "Int4 quantization OOM is a memory/offload problem; tokenization changes are a separate behavior change." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Both involve Qwen3-family multimodal models, but one is a video-input ValueError and the other is a weight_only loading error." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to the T5 apex FusedRMSNorm memory leak." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Void segmentation-map processing and BarkProcessor voice_preset handling are unrelated processor bugs." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "torch.compile + SDPA on Bamba is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both mention loading/offload, but one is CPU offload with quantization and the other is BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:44534", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffer initialization junk and from_pretrained inside Python classes are distinct issues." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3 auto device mapping cudaErrorAssert and GraniteSpeech PEFT local-checkpoint loading are different failures." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoModelForCausalLM.from_config text_config unwrapping are not the same bug." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A cache-strategies typo in docs is unrelated to stale docs mentioning removed pipeline tasks." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming chat/completions failures are different components and symptoms." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Trailing-slash local model import warnings are unrelated to the multimodal video-branch undefined-variable bug." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Metric gathering truncation and SAM3 text/bbox weird behavior are separate issues." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "CPU offload/quantization device mismatch is unrelated to tokenizer special-token registration." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token-count tracking and progressive generation are unrelated." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor preset handling and Qwen3 auto device mapping/cudaErrorAssert are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:46821", + "accept": false, + "reason": "Qwen3-MoE sliding_window and TimesFM window_size AttributeError are different models and APIs." + }, + { + "left": "issue:43540", + "right": "issue:46858", + "accept": false, + "reason": "Qwen3OmniMoe video ValueError and static-cache seq_length tensor arithmetic are unrelated." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Qwen3.5/Qwen3.6 config unwrapping and CPU-offload rotary embedding compatibility are different problems." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Incorrect visual-encoder keys on save_pretrained and missing text-encoder weight loading are different load/save regressions." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Loading ScandiBERT and tokenizer behavior changes across versions are not the same underlying issue." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Optional post_init policy for patchtsmixer and Qwen3.5 num_labels propagation are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding_window and Float8 storage lookup errors are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:46665", + "accept": false, + "reason": "Non-persistent buffer junk in Transformers 5.0 and from_pretrained-in-class behavior are unrelated." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken special-argument duplication and ScandiBERT loading failure are different tokenizer/model-loading issues." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver interpolation on non-default resolution and bfloat16 dtype mismatches are distinct bugs." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 and degenerate zero loss in sequence classification are not the same code path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 17, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43295", + "issue:43381", + "issue:43421", + "issue:43493", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44186", + "issue:44368", + "issue:44514", + "issue:44545", + "issue:44610", + "issue:44671", + "issue:44843", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:45042", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45507", + "issue:45704", + "issue:45759", + "issue:45834", + "issue:45907", + "issue:45923", + "issue:45987", + "issue:46032", + "issue:46050", + "issue:46463", + "issue:46655", + "issue:46693", + "issue:46829", + "issue:46832", + "issue:46858" + ], + "soft_pairs": [ + "issue:43381|issue:43844", + "issue:44987|issue:45042", + "issue:44368|issue:45357", + "issue:39692|issue:44610", + "issue:41720|issue:45704", + "issue:43746|issue:46050", + "issue:43421|issue:43927", + "issue:46655|issue:46693", + "issue:43381|issue:45704", + "issue:45704|issue:45923", + "issue:42886|issue:44843", + "issue:44918|issue:44936", + "issue:45759|issue:45923", + "issue:43746|issue:43873", + "issue:43421|issue:43653", + "issue:43994|issue:44671", + "issue:44514|issue:44545", + "issue:46032|issue:46463", + "issue:39692|issue:43493", + "issue:43421|issue:44186", + "issue:45081|issue:45356", + "issue:43782|issue:45381", + "issue:30064|issue:36010", + "issue:45507|issue:46858", + "issue:45356|issue:45907", + "issue:39692|issue:43873", + "issue:43720|issue:43819", + "issue:43653|issue:43720", + "issue:46829|issue:46832", + "issue:45987|issue:46829", + "issue:43295|issue:45834" + ], + "prepared_review_unit_hash": "1ce66ff6d2642937239b05db676cb75233e915188c7b47cd58ca8bb8d9d9161f", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37450, + "estimated_input_tokens": 9363, + "estimated_eval_tokens": 18982 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous transformer/model bugs with one clear duplicate pair and one likely offline-tokenizer duplicate. No PRs are present.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44545 is the cleanest duplicate target for the only exact issue duplication in the set (same Qwen2_5_VLProcessor.apply_chat_template crash as issue:44514).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44545 is the best overall cluster anchor: it is narrowly scoped, clearly reproducible, and has the strongest duplicate signal; the rest of the items are largely unrelated sub-issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Different bugs: eval-mode gradient checkpointing misuse vs ZeRO-3/deepspeed gradient inflation." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Unrelated failures: a model-loading regression vs PIL image processors incorrectly requiring torchvision." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "Different code paths and symptoms: a tie_word_embeddings warning during LoRA fine-tuning vs incorrect visual encoder keys in save_pretrained." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Different models and defects: SigLIP2 doc example errors vs OmDet-Turbo processor/model input-size mismatch." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "No shared bug: Qwen3 device-mapping CUDA assert is unrelated to T5/apex RMSNorm memory leak." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "Different loading/save problems: PEFT adapter local-checkpoint loading vs sinq quantized model save/load breakage." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both involve tokenizers, but the failures differ: runtime post-processor updates vs config persistence losing custom token IDs." + }, + { + "left": "issue:46655", + "right": "issue:46693", + "accept": false, + "reason": "Different subsystems: video-processor padding math vs FA2 long-context generation performance regression." + }, + { + "left": "issue:43381", + "right": "issue:45704", + "accept": false, + "reason": "Completely different bugs in different components." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "Different issues: apex RMSNorm leak vs missing supports_gradient_checkpointing flag." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": true, + "reason": "Same underlying offline tokenizer failure: both describe HF_HUB_OFFLINE cache-loading breakage, and 44843 identifies the unconditional model_info() call that explains 42886." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Different trainer failures: SFT unpacking embeddings vs evaluate() failing after train()." + }, + { + "left": "issue:45759", + "right": "issue:45923", + "accept": false, + "reason": "Different model/config problems: unwrapping text_config vs missing gradient checkpointing flag." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "Both mention loading/quantization, but one is PEFT checkpoint loading and the other is offloading with quantization; not the same concrete bug." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Related tokenizer area, but not the same bug: post-processor update logic vs BigBirdTokenizer special-token registration/empty decode." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different models and symptom classes: SigLIP2 inference mismatch vs CamemBERT masked-LM prediction regression." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Exact duplicate report: same Qwen2_5_VLProcessor.apply_chat_template batched-input crash with padding=False." + }, + { + "left": "issue:46032", + "right": "issue:46463", + "accept": false, + "reason": "Unrelated bugs: Mamba2Mixer cache correctness vs T5 decoder input error-message wording." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Same model family, but different problems: doc/example misuse and quantization failure vs a core HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Both are tokenizer issues, but the specific failures differ: special-token post-processor updates vs NER/batched padding-truncation crashes." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both reference Mistral regex handling, but the reported failures are different enough to treat as separate regressions." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen VL bugs: weight_only loading error vs incorrect vision_position_ids for video input." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "No overlap: image-processing segmentation bug vs GenerationMixin import error." + }, + { + "left": "issue:45507", + "right": "issue:46858", + "accept": false, + "reason": "Different code paths: invalid method call in GraniteMoEHybrid vs static-cache seq-length tensor breaking slice arithmetic." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Unrelated: tokenizer regression in Kimi-K2.5 vs list_repo_files not forwarding revision." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "Different problems: SigLIP2 docs/example failures vs quantization/offloading behavior." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Different model implementations and failures: BitNet packed-weight loading vs DAC from_latents/STE mismatch." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "No shared underlying bug: BigBirdTokenizer special-token handling vs BitNet accelerate-loading of packed weights." + }, + { + "left": "issue:46829", + "right": "issue:46832", + "accept": false, + "reason": "Different vision-model bugs: incorrect hidden_states output vs RT-DETR feature-level indexing error." + }, + { + "left": "issue:45987", + "right": "issue:46829", + "accept": false, + "reason": "Cache API return-type bug vs VideoPrism output-field misassignment." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Both are multimodal/processor issues, but they affect different models and different failure modes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly identifies the exact duplicate pair (44514/44545), the likely offline-tokenizer duplicate (42886/44843), and does not claim any PRs exist. The soft-edge verdicts are consistently conservative, with no clear overstatement of evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 18, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:30064", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43335", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43596", + "issue:43611", + "issue:43618", + "issue:43643", + "issue:43644", + "issue:43653", + "issue:43697", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43927", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44361", + "issue:44442", + "issue:44448", + "issue:44488", + "issue:44661", + "issue:44671", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45834", + "issue:45987", + "issue:46294", + "issue:46665", + "issue:46693", + "issue:46858", + "issue:46932" + ], + "soft_pairs": [ + "issue:43550|issue:43927", + "issue:44062|issue:44488", + "issue:43550|issue:44361", + "issue:43540|issue:46294", + "issue:45381|issue:46693", + "issue:43531|issue:43643", + "issue:43596|issue:44220", + "issue:43611|issue:46665", + "issue:44008|issue:46858", + "issue:44186|issue:44898", + "issue:43653|issue:43819", + "issue:41720|issue:44991", + "issue:45987|issue:46932", + "issue:30064|issue:36331", + "issue:43643|issue:44877", + "issue:44220|issue:44661", + "issue:43618|issue:43697", + "issue:43335|issue:43756", + "issue:43550|issue:43819", + "issue:44843|issue:45081", + "issue:44448|issue:44671", + "issue:39692|issue:45834", + "issue:44186|issue:44442", + "issue:44821|issue:45084", + "issue:43540|issue:44008", + "issue:34634|issue:36010", + "issue:45183|issue:45479", + "issue:43644|issue:43749" + ], + "prepared_review_unit_hash": "c77ddd8870b9457b2fe9f7e5d82042246e1b847445dee1d15456d24263fdd32b", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36259, + "estimated_input_tokens": 9065, + "estimated_eval_tokens": 18386 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set: the issues span unrelated bugs across tokenizers, vision/video models, training, config loading, and cache behavior. No safe canonical issue/PR stands out.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits this cluster; the reports describe distinct failures in different components and code paths.", + "canonical_pr_reason": "No PRs are present in the provided items.", + "best_issue_reason": "No global best issue candidate: the items are heterogeneous and not interchangeable as a representative duplicate target.", + "best_pr_reason": "No PRs are present in the provided items.", + "soft_edge_verdicts": [ + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Different failures: torch.compile+SDPA on Bamba vs custom token IDs lost after save/load on DiaConfig." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "Both are tokenizer-related, but one is an AddedToken keyword conflict and the other is a model-specific load failure." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Unrelated models and root causes: Bamba compile/SDPA vs MLukeTokenizer AttributeError." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both involve video inputs, but they affect different Qwen models and different mismatch logic." + }, + { + "left": "issue:45381", + "right": "issue:46693", + "accept": false, + "reason": "Different subsystems: Qwen2.5-VL video position IDs vs FA2 long-context performance regression." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to AutoConfig missing fields with trust_remote_code." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "DeepSpeed ZeRO3/BertModel index error is unrelated to the fbank feature extraction issue." + }, + { + "left": "issue:43611", + "right": "issue:46665", + "accept": false, + "reason": "Both mention from_pretrained, but the concrete bugs differ: base_model_prefix loading break vs class-context inconsistency." + }, + { + "left": "issue:44008", + "right": "issue:46858", + "accept": false, + "reason": "Gemma3n forward variable collision is unrelated to static cache seq_length tensor behavior." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Different models and failures: LayoutLMv2Tokenizer input handling vs Perceiver image classification resolution handling." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token registration bug is unrelated to DAC from_latents vs forward mismatch." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:45987", + "right": "issue:46932", + "accept": false, + "reason": "StaticCache seq_length contract bug is unrelated to Gemma4 quantized vision/audio dtype casting." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "ImageProcessor segmentation-map failure is unrelated to CustomTrainer compute_loss argument mismatch." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-loading issues, but one is missing fields with trust_remote_code and the other is strict config rejection for granite_speech." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio fbank feature extraction and add-new-model-like TOKENIZER_MAPPING_NAMES handling are separate bugs." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Different output regressions in different models; no shared concrete code-path problem is evident." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and Smollm3 RoPE-layer dropping are unrelated architecture bugs." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "torch.compile+SDPA failure on Bamba is unrelated to DAC from_latents/forward mismatch." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Same helper area (_patch_mistral_regex), but different concrete failures: offline model_info call vs backend_tokenizer attribute error." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Both are v5 output regressions, but they affect different models and likely different code paths." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 doc example errors are unrelated to Kosmos2.5 long-OCR index error." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2Tokenizer crash is unrelated to AutoTokenizer failing to load FastSpeech2ConformerTokenizer." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and template-node compilation are unrelated problems." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Different Gemma/Qwen code paths and failure modes; both are multimodal but not the same bug." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset and GenerationMixin import error are unrelated issues." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Both are multi-model behavior changes, but one is input validation messaging and the other is classification loss degeneracy." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer junk in Transformers 5.0.0 is unrelated to broken FSDP CPU RAM efficient loading." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 19, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:34634", + "issue:36331", + "issue:43322", + "issue:43388", + "issue:43421", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43645", + "issue:43716", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43818", + "issue:43825", + "issue:43827", + "issue:43931", + "issue:43994", + "issue:44008", + "issue:44291", + "issue:44479", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44811", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45381", + "issue:45584", + "issue:45685", + "issue:45701", + "issue:45834", + "issue:46032", + "issue:46291", + "issue:46294", + "issue:46463", + "issue:46512", + "issue:46665", + "issue:46846" + ], + "soft_pairs": [ + "issue:45584|issue:45834", + "issue:43322|issue:45685", + "issue:43606|issue:45072", + "issue:43388|issue:46032", + "issue:43825|issue:43994", + "issue:44479|issue:45834", + "issue:43550|issue:45072", + "issue:43782|issue:44977", + "issue:43421|issue:44811", + "issue:44561|issue:45020", + "issue:43577|issue:43818", + "issue:46294|issue:46463", + "issue:44991|issue:46291", + "issue:43582|issue:46512", + "issue:43716|issue:44008", + "issue:43931|issue:46846", + "issue:28218|issue:29159", + "issue:44805|issue:44936", + "issue:44610|issue:45072", + "issue:43749|issue:44291", + "issue:43540|issue:45381", + "issue:43720|issue:45072", + "issue:43825|issue:43827", + "issue:43645|issue:46665", + "issue:44671|issue:45701", + "issue:43577|issue:43582", + "issue:34634|issue:36331" + ], + "prepared_review_unit_hash": "501d8f0033fcf82ebecae9f59cbd77516c424850ba632712de69269d3feea9f4", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35468, + "estimated_input_tokens": 8867, + "estimated_eval_tokens": 17990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated issue reports across tokenizer, multimodal, loading, dtype, and trainer subsystems. I found no convincing duplicate pairs, so this cluster should not be merged into a single canonical bug.", + "confidence": 0.97, + "canonical_issue_reason": "No true duplicate stands out; if one issue must anchor the cluster, issue 45701 is the broadest tokenizer-regression report and the closest to the small tokenizer-related overlap, but the set is still too heterogeneous.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45701 is the most general and potentially umbrella-like of the listed issues, but only by breadth; it is not a strong duplicate canonical because the rest of the cluster spans many distinct failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Different bugs: empty transcription / special-token alignment vs long OCR input index handling." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Unrelated failures: Llava Next load segfault vs MPS histogram TypeError on Int tensors." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Both involve loading/device-type problems, but the concrete failures and models differ too much to be duplicates." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Metric gathering label truncation and Mamba2 cache corruption are separate code paths." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "An error-message wording issue is not the same as a model producing nonsensical outputs." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Different model families and different failure modes: video regression vs OCR index error." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "torch.compile+SDPA failure is distinct from bf16 dtype mismatch." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Weight-only loading error and flash-attention generation bug are not the same underlying issue." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Both are tokenizer/processor related, but one is post-processor sync at runtime and the other is batch_decode skip_special_tokens handling." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Both mention trust_remote_code, but one is a missing helper removal and the other is a broader version regression." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype persistence and Video-LLaVA temporal-attention/weight-sharing problems are unrelated." + }, + { + "left": "issue:46294", + "right": "issue:46463", + "accept": false, + "reason": "Qwen video chat-template mismatch and T5 decoder validation messaging are separate issues." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Tokenizer loading failure and kernels dependency import failure are different regressions." + }, + { + "left": "issue:43582", + "right": "issue:46512", + "accept": false, + "reason": "Apple Silicon allocator TypeError and stale model.config.dtype after from_config are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "dtype mismatch in preprocessing vs variable-name collision AttributeError are not the same bug." + }, + { + "left": "issue:43931", + "right": "issue:46846", + "accept": false, + "reason": "Both are Qwen3-VL/Qwen3.5 loading issues, but one is a weight-shape mismatch and the other is a TP plan/OOM problem." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "Both are tokenizer complaints, but one is added-token spacing and the other is slow/fast tokenizer inconsistency; not enough evidence of a single root cause." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask-shape IndexError and train-then-evaluate failure do not describe the same code-path problem." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Image-size mismatch in a processor is unrelated to bf16 dtype mismatch." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP RAM-efficient loading and init_empty_weights _is_hf_initialized argument errors are distinct loading regressions." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both are video-input bugs in Qwen models, but they target different models and different failure mechanisms." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading is a different issue from dtype mismatches." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Incorrect error messaging and outdated docs references are separate changes." + }, + { + "left": "issue:43645", + "right": "issue:46665", + "accept": false, + "reason": "Notebook custom-model initialization and from_pretrained inside Python classes are related only at a high level, not the same bug." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM regression is not the same as a generic tokenization change report." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "BLIP2 dtype state and Apple Silicon allocator warmup are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 20, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:42175", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43582", + "issue:43638", + "issue:43645", + "issue:43653", + "issue:43697", + "issue:43742", + "issue:43782", + "issue:43819", + "issue:43906", + "issue:43927", + "issue:44060", + "issue:44368", + "issue:44554", + "issue:44589", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44991", + "issue:45003", + "issue:45020", + "issue:45042", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45372", + "issue:45406", + "issue:45538", + "issue:45600", + "issue:46175", + "issue:46209", + "issue:46285", + "issue:46291", + "issue:46326", + "issue:46697", + "issue:46728", + "issue:46932" + ], + "soft_pairs": [ + "issue:42175|issue:45081", + "issue:46175|issue:46932", + "issue:45003|issue:46285", + "issue:46697|issue:46728", + "issue:43653|issue:44912", + "issue:43782|issue:45325", + "issue:43262|issue:43819", + "issue:45372|issue:46326", + "issue:45081|issue:45406", + "issue:43582|issue:44857", + "issue:43742|issue:45020", + "issue:44589|issue:45042", + "issue:43550|issue:44912", + "issue:43452|issue:44991", + "issue:43257|issue:43329", + "issue:43526|issue:43697", + "issue:43329|issue:45600", + "issue:44060|issue:44368", + "issue:43906|issue:46209", + "issue:36010|issue:36331", + "issue:43526|issue:44554", + "issue:36010|issue:43645", + "issue:45084|issue:45538", + "issue:43819|issue:43927", + "issue:43638|issue:44936", + "issue:43531|issue:45042", + "issue:42175|issue:46291", + "issue:42490|issue:43257" + ], + "prepared_review_unit_hash": "1333573b8ce8c764c792deb91c7954ab829689b07aa824ee1d20170f25a78369", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 36025, + "estimated_input_tokens": 9007, + "estimated_eval_tokens": 18270 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly independent Transformers issues spanning unrelated subsystems; none of the soft pairs look like true duplicates or mergeable PR-equivalents.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue: the cluster mixes unrelated bugs across import, loading, tokenizer, processor, model, and training paths, so one issue would not represent the group well.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue stands out; the items are heterogeneous, so any single issue would be an arbitrary representative rather than a true canonical.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Different problems: backend packaging/dependency selection versus a Mistral regex patch crash during tokenizer loading." + }, + { + "left": "issue:46175", + "right": "issue:46932", + "accept": false, + "reason": "Both are Gemma4-related, but one is a huge one-hot materialization issue and the other is quantized embedder dtype casting; different code paths." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "sys.modules access in modeling_utils is unrelated to the CLIP text-model structure regression." + }, + { + "left": "issue:46697", + "right": "issue:46728", + "accept": false, + "reason": "Different model processors and symptoms: Swin2SR over-padding versus Aria patch-count miscalculation." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token registration bug versus MXFP4 quantization fallback; different subsystems." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3VL weight-only loading error is separate from Qwen2.5-VL rope index scaling." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting is unrelated to DAC latent decoding mismatch." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Both involve mistral_common, but one is a missing import breaking Gemma4 loading and the other is a continue_final_message tokenizer bug." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Different failures: tokenizer regex patch crashes versus Gemma4Processor missing _tokenizer in serve." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon TypeError in caching warmup is unrelated to CUDA AMP loss crashes." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "A specific MobileLLM loading key error is not the same as the broader remote_code regression." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to PIL image processors requiring torchvision." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile SDPA failure on Bamba is a different bug from MXFP4 quantized loading fallback." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "gguf_file handling for from_pretrained is unrelated to the est-roberta tokenizer load regression." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed is different from an undefined-variable bug in multimodal token counting." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "Beit reduce_labels returning one label is unrelated to RTDetrV2 output drift." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Undefined video-branch variables in multimodal token counting do not match removed auto-mapping config references." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings, but one is an incorrect internal tie and the other is a LoRA/config warning about tie_word_embeddings." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated reproduction of another issue is not the same as the FP8 inference failure on B200." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "GenerationMixin import failure and CustomTrainer compute_loss signature mismatch are distinct API breakages." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "ImageProcessor label reduction bug is unrelated to the MPS attention correctness issue." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "ImportError for GenerationMixin is not the same as the Jupyter custom-model initialization breakage." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template compilation TypeError and CLIPTokenizer model_max_length are unrelated tokenizer/model config issues." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC forward/from_latents mismatch is unrelated to DiaConfig losing custom token IDs after save/load." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "deepspeed zero3 IndexError with a non-pretrained Bert model is different from trainer.evaluate() failing after train()." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to PIL backend torchvision requirements." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Package backend selection for torch installs is unrelated to kernels v0.15 breaking pipelines import." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Trailing-slash local-path import warning is unrelated to Qwen3 MoE weight conversion under accelerate/deepspeed." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 21, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28282", + "issue:29159", + "issue:29870", + "issue:34567", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:42915", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43454", + "issue:43540", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43650", + "issue:43696", + "issue:43720", + "issue:43742", + "issue:43746", + "issue:43749", + "issue:43854", + "issue:43906", + "issue:43927", + "issue:43931", + "issue:44291", + "issue:44387", + "issue:44479", + "issue:44568", + "issue:44610", + "issue:44617", + "issue:44756", + "issue:44811", + "issue:44821", + "issue:44912", + "issue:44977", + "issue:45042", + "issue:45356", + "issue:45459", + "issue:45759", + "issue:45803", + "issue:45859", + "issue:46421", + "issue:46424", + "issue:46701" + ], + "soft_pairs": [ + "issue:43720|issue:44610", + "issue:43746|issue:44811", + "issue:43854|issue:44977", + "issue:43643|issue:45759", + "issue:39692|issue:44912", + "issue:38175|issue:44610", + "issue:43257|issue:43696", + "issue:43262|issue:44811", + "issue:43257|issue:43931", + "issue:44387|issue:44617", + "issue:43329|issue:43450", + "issue:43329|issue:44479", + "issue:45459|issue:45803", + "issue:46421|issue:46424", + "issue:34567|issue:40990", + "issue:42915|issue:43742", + "issue:43749|issue:44756", + "issue:44291|issue:45042", + "issue:29159|issue:34567", + "issue:43650|issue:43906", + "issue:43606|issue:45859", + "issue:42915|issue:44821", + "issue:43540|issue:46701", + "issue:43927|issue:44912", + "issue:28282|issue:29870", + "issue:44568|issue:45356", + "issue:43454|issue:43577" + ], + "prepared_review_unit_hash": "a220334d26285d2dd06c841dad84970622cdf190cbf1205602e0bba569a4c6c5", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35448, + "estimated_input_tokens": 8862, + "estimated_eval_tokens": 17980 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The items are a heterogeneous mix of unrelated bug reports across loading, tokenization, multimodal processing, quantization, and memory issues. None of the proposed soft pairs look like the same underlying bug/change, so no duplicate merges are justified.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43749 is the strongest single anchor in the set: it has the highest discussion/inbound-reference signal and describes a concrete loading-path regression. That said, the overall set is not a true duplicate cluster, so this is only a pragmatic representative.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43749 is the best issue to surface as the representative artifact because it is well-scoped, high-signal, and more broadly actionable than the other reports in this mixed set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Different failures: BitNet packed-weight unpacking during accelerate loading vs OmDet-Turbo processor image-size mismatch." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "Different code paths: PEFT adapter checkpoint loading vs Whisper processor batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms: GLM-4.7 unit-test loading failure vs Qwen3.5 flash-attention generation issue." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Both involve config/from_pretrained behavior, but one is missing fields with trust_remote_code and the other is AutoModelForCausalLM.from_config not unwrapping text_config for composite Qwen multimodal configs." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Different topics: SigLIP2 doc example errors vs gpt-oss-20b MXFP4 quantization fallback." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Different bug classes: unexpected zero probabilities in SigLIP2 vs processor/model input-size mismatch in OmDet-Turbo." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MoE weight conversion with accelerate+deepspeed is unrelated to GPT-oss-20b CUDA OOM." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sampling-rate default mismatch is unrelated to Whisper batch_decode special-token skipping." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen-related loading problems, but one is MoE weight conversion under accelerate/deepspeed and the other is a weight-shape mismatch for Qwen3-VL-30B." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Different memory symptoms and causes: int4 reserved-memory regression in Transformers 5.x vs Sam3Video CUDA OOM." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Video-related only at a broad level; one is undefined variables in multimodal token counting, the other is incorrect batched video processor output shape." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "Both touch video/VLM processing, but the bugs are distinct: undefined helper/merge_size in token counting vs a v5.3.0 video-input regression across several models." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both concern error handling, but one is protobuf decode error masking in tokenizers and the other is a bare except in Fuyu tensor conversion." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both mention cache behavior, but one is lazy initialization causing recompilation and the other is max_cache_length being honored only for Quantized Cache." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Training token-count tracking bug vs GPT-oss-20b high-perplexity evaluation issue; unrelated failures." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Both involve model loading/configuration, but one is Qwen3Moe FineGrainedFP8Config failure and the other is a MobileLLM key error." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Both are memory-related, but one is FSDP_CPU_RAM_EFFICIENT_LOADING breakage and the other is mmap causing OOM on Strix Halo." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Different regression types: init_empty_weights/_is_hf_initialized TypeError vs PIL backend processors wrongly requiring torchvision." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState token-count updates." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "No concrete shared bug: one title is non-informative, the other is an isolated reproduction of a previously reported issue." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both mention CPU/offload-like behavior, but one is a device mismatch in bark-small CI and the other is a Qwen3_5Moe rotary embedding CPU-offload incompatibility." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure is unrelated to AutoImageProcessor URL loading." + }, + { + "left": "issue:43540", + "right": "issue:46701", + "accept": false, + "reason": "Different multimodal/video pipeline bugs: Qwen3OmniMoe video input ValueError vs DiffusionGemmaGenerationOutput pipeline failure." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "DiaConfig custom-token persistence bug is unrelated to gpt-oss-20b MXFP4 quantization fallback." + }, + { + "left": "issue:28282", + "right": "issue:29870", + "accept": false, + "reason": "ImportError for missing PyTorch is unrelated to Bart generate() decoder_attention_mask unbound-variable error." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Different tokenizer regressions: add_special_tokens not adding BOS/EOS vs Kimi-K2.5 codec-handling and warning regression." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision lm_head weight tying bug is unrelated to BLIP2 dtype propagation to model and qformer." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 22, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29127", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:36331", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43404", + "issue:43425", + "issue:43452", + "issue:43525", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43824", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44560", + "issue:44568", + "issue:44849", + "issue:44857", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45127", + "issue:45137", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45488", + "issue:45538", + "issue:45701", + "issue:45907", + "issue:46291", + "issue:46512", + "issue:46693", + "issue:46821", + "issue:46832", + "issue:46869" + ], + "soft_pairs": [ + "issue:44987|issue:45538", + "issue:28282|issue:34689", + "issue:43525|issue:46291", + "issue:43697|issue:45357", + "issue:33357|issue:34567", + "issue:44857|issue:46832", + "issue:43404|issue:45127", + "issue:43425|issue:46291", + "issue:45538|issue:46869", + "issue:45137|issue:45405", + "issue:42915|issue:45538", + "issue:42175|issue:43824", + "issue:28282|issue:29127", + "issue:44479|issue:46693", + "issue:44062|issue:44991", + "issue:43742|issue:44291", + "issue:42915|issue:43749", + "issue:44936|issue:46821", + "issue:24643|issue:29159", + "issue:43452|issue:43582", + "issue:43577|issue:46512", + "issue:44849|issue:45701", + "issue:44987|issue:45907", + "issue:27957|issue:28282", + "issue:41720|issue:43540", + "issue:43540|issue:44560", + "issue:44479|issue:45381", + "issue:44568|issue:45488", + "issue:34567|issue:36331" + ], + "prepared_review_unit_hash": "7cf32ece9642dcf00d943c0e0d1e66a753ec1609fd039cc811642ac0cde5da91", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36367, + "estimated_input_tokens": 9092, + "estimated_eval_tokens": 18440 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue set. One likely duplicate pair exists around the Qwen2.5-VL video regression (44479/45381); the rest of the soft matches are distinct bugs across different models/components and should stay separate.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44479 is the best representative of the only accepted duplicate pair: it is the broader, earlier report for the Qwen video-input regression, while issue:45381 is a narrower symptom of the same 5.3.0 bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 most clearly captures the shared Qwen2.5-VL video regression and is the most general surviving issue in the only real duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Different failures: model loading regression vs CLIPTokenizer max-length bug." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Both are loading errors, but they concern different models and different root causes." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Unrelated issues: missing pad_token_id on Llama4Config vs kernels revision/version validation." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "Different problems: RTDetrV2 output changes vs Qwen3.5 save_pretrained key serialization." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error in CLIP loading is unrelated to TrainerState.num_input_tokens_seen not updating." + }, + { + "left": "issue:44857", + "right": "issue:46832", + "accept": false, + "reason": "Related RT-DETR family, but different concrete bugs and code paths." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "One is a weight-tying bug; the other is LoRA merge collapse with extended vocab." + }, + { + "left": "issue:43452", + "right": "issue:46291", + "accept": false, + "reason": "Different subsystems: GGUF file handling vs pipelines import-time kernel revision checks." + }, + { + "left": "issue:45538", + "right": "issue:46869", + "accept": false, + "reason": "CLIPTokenizer max_length bug is unrelated to YOLO multi-GPU device-map failures." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error is unrelated to the PEFT minimum-version bump." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Qwen3Moe FP8 failure and CLIPTokenizer max_length are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Packaging/backend dependency issue vs missing Qwen2_5_VL import are different bugs." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "AutoModel PyTorch import error is unrelated to the layoutlmv3 box-validation message quality issue." + }, + { + "left": "issue:44479", + "right": "issue:46693", + "accept": false, + "reason": "Both are regressions, but one is a Qwen video-input bug and the other is a FlashAttention2 long-context performance regression." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "AddedToken keyword collision is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Different loading failures: MobileLLM key error vs init_empty_weights unexpected _is_hf_initialized argument." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is unrelated to FSDP CPU RAM efficient loading." + }, + { + "left": "issue:44936", + "right": "issue:46821", + "accept": false, + "reason": "Trainer evaluate/train interaction is unrelated to TimesFM window_size attribute errors." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed training weight-shape error and slow/fast tokenizer inconsistency are different bugs." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF loading and AppleSilicon caching_allocator_warmup TypeError do not share the same failure path." + }, + { + "left": "issue:43577", + "right": "issue:46512", + "accept": false, + "reason": "Both involve dtype, but one is an actual Blip2 model-casting bug and the other is stale config.dtype after from_config." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-states bug is unrelated to a generic tokenization-version change report." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Model-loading failure is unrelated to list_repo_files needing to forward revision." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 behavior is unrelated to AutoModel PyTorch import errors." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different Qwen failures: auto device-map cudaErrorAssert vs video-input ValueError in Qwen3OmniMoe." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both mention video inputs, but they are different models and distinct error paths." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Same underlying Qwen2.5-VL video-input regression in transformers 5.3.0; 45381 is a narrower symptom of 44479." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Tokenizer special-token addition bug and LlamaTokenizer ByteLevel-to-Metaspace override are different regressions." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token-count update bug is unrelated to CustomTrainer.compute_loss signature change." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it identifies the single accepted duplicate-like pair (44479/45381) using matching Qwen video-input regression language, and the remaining soft rejections are phrased as distinct bugs with different models/components. I don\u2019t see evidence of overclaiming or an unsafe mergeable-pair assertion." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 23, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:24643", + "issue:29870", + "issue:38175", + "issue:42175", + "issue:43257", + "issue:43452", + "issue:43531", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43716", + "issue:43742", + "issue:43756", + "issue:43906", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44737", + "issue:44792", + "issue:44821", + "issue:44877", + "issue:44898", + "issue:44964", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45405", + "issue:45600", + "issue:45663", + "issue:45672", + "issue:45674", + "issue:45823", + "issue:46399", + "issue:46463", + "issue:46468", + "issue:46491", + "issue:46693", + "issue:46701", + "issue:46821", + "issue:46829", + "issue:46846" + ], + "soft_pairs": [ + "issue:44792|issue:44964", + "issue:38175|issue:44898", + "issue:42175|issue:44488", + "issue:43257|issue:43756", + "issue:43540|issue:44479", + "issue:44062|issue:45081", + "issue:44291|issue:44821", + "issue:45042|issue:45081", + "issue:43756|issue:46399", + "issue:43756|issue:45325", + "issue:43582|issue:46821", + "issue:43742|issue:44589", + "issue:43716|issue:46701", + "issue:44568|issue:46491", + "issue:45405|issue:45600", + "issue:45672|issue:45674", + "issue:45823|issue:46463", + "issue:43906|issue:46468", + "issue:43577|issue:45072", + "issue:43531|issue:44821", + "issue:43906|issue:46693", + "issue:44737|issue:46463", + "issue:24643|issue:29870", + "issue:43452|issue:44488", + "issue:44625|issue:44877", + "issue:45663|issue:46846", + "issue:43697|issue:46829", + "issue:43540|issue:46693" + ], + "prepared_review_unit_hash": "9c22aef4c3995f45aad7ad369678473bec9327fd319a96411612f51c67925c33", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35882, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated bug reports spanning model loading, tokenizer/config edge cases, multimodal/video regressions, and quantization/runtime errors. I don\u2019t see any soft pair that clearly describes the same concrete underlying bug or fix.", + "confidence": 0.94, + "canonical_issue_reason": "issue:44479 is the strongest representative: it describes a concrete regression, names multiple affected models, and has the clearest scope/version framing among the set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the best single issue to keep as the cluster representative because it is specific, user-facing, and documents a broader regression pattern than most others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Different models and failures: a janus image-generation test issue vs a Phi-4 multimodal load failure." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision-model bugs, but the symptoms and code paths differ: zero probabilities in SigLIP2 vs Perceiver resolution/interpolation failure." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Unrelated: packaging/backend dependency omission vs a specific model loading failure for cjvt/sleng-bert." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Different underlying problems: DeepSpeed/accelerate weight conversion vs Smollm3 RoPE layer dropping." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Similar video-input theme, but the affected model families and reported breakage differ enough that they do not look like the same bug." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer-related, but different failures: AddedToken argument duplication vs Mistral regex patch hitting a missing backend_tokenizer attribute." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "Different loading surfaces and errors: init_empty_weights/_is_hf_initialized vs AutoImageProcessor-from-URL handling." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "One is an image processor backend dependency regression; the other is a tokenizer regex patch crash." + }, + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Both touch Qwen-related model internals, but one is RoPE layer selection and the other is kernel registration/decorator handling." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen-family rotary-position bugs affecting different models and logic paths." + }, + { + "left": "issue:43582", + "right": "issue:46821", + "accept": false, + "reason": "Completely different subsystems: Apple Silicon allocator warmup vs TimesFM window_size attribute handling." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated load-time issues: config key error for MobileLLM vs missing Float8 storage type." + }, + { + "left": "issue:43716", + "right": "issue:46701", + "accept": false, + "reason": "Both are multimodal, but one is dtype mismatch in Mistral-3 preprocessing and the other is a DiffusionGemma pipeline failure." + }, + { + "left": "issue:44568", + "right": "issue:46491", + "accept": false, + "reason": "Both are tokenizer regressions, but they involve different tokenizers and different behaviors (special tokens vs encode-decode spacing)." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "Unrelated maintenance issues: PEFT version pin vs stale auto_mappings config references." + }, + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Both concern 4-bit inference, but the concrete bugs differ: uint8/torch.finfo in Gemma4 audio modules vs llm_int8_skip_modules clearing lm_head exclusion." + }, + { + "left": "issue:45823", + "right": "issue:46463", + "accept": false, + "reason": "Different problems entirely: Gemma4 device placement vs T5 decoder-input validation messaging." + }, + { + "left": "issue:43906", + "right": "issue:46468", + "accept": false, + "reason": "No shared bug: one is a reproduction of an existing issue, the other is an official Nemotron example breakage." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype, but one is Blip2 dtype propagation and the other is bfloat16 inference mismatches in different models." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Different model families and failures: Qwen3-MoE sliding window vs AutoImageProcessor URL loading." + }, + { + "left": "issue:43906", + "right": "issue:46693", + "accept": false, + "reason": "A reproduction of an old issue is not the same as a new FA2 long-context performance regression." + }, + { + "left": "issue:44737", + "right": "issue:46463", + "accept": false, + "reason": "XLNet CPU device-placement bug vs T5 validation error message; no shared code-path." + }, + { + "left": "issue:24643", + "right": "issue:29870", + "accept": false, + "reason": "Different training-time failures: DeepSpeed weight shape error vs Bart generate decoder_attention_mask initialization bug." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading issues, but one is a gguf/from_pretrained integration break and the other is a specific model load failure." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Different config problems: Qwen3.5 num_labels propagation vs strict config blocking granite_speech loading." + }, + { + "left": "issue:45663", + "right": "issue:46846", + "accept": false, + "reason": "Both are Gemma4-related, but one is FSDP2 shared_kv_states rebuilding and the other is missing tensor-parallel plan entries." + }, + { + "left": "issue:43697", + "right": "issue:46829", + "accept": false, + "reason": "Different output-shape/output-field bugs in different models; not the same concrete issue." + }, + { + "left": "issue:43540", + "right": "issue:46693", + "accept": false, + "reason": "One is a video-input ValueError for a Qwen multimodal model, the other a long-context FA2 performance regression; same broad area, different bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 24, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28282", + "issue:29159", + "issue:32090", + "issue:33357", + "issue:38175", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43525", + "issue:43526", + "issue:43643", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43749", + "issue:43761", + "issue:43819", + "issue:43994", + "issue:44361", + "issue:44442", + "issue:44466", + "issue:44493", + "issue:44737", + "issue:44811", + "issue:44843", + "issue:44898", + "issue:44913", + "issue:44936", + "issue:45005", + "issue:45200", + "issue:45216", + "issue:45356", + "issue:45357", + "issue:45372", + "issue:45375", + "issue:45479", + "issue:45588", + "issue:45920", + "issue:46050", + "issue:46121", + "issue:46170", + "issue:46463", + "issue:46701", + "issue:46829", + "issue:46832", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:45357|issue:45920", + "issue:44913|issue:46121", + "issue:28282|issue:32090", + "issue:46463|issue:46899", + "issue:44737|issue:45588", + "issue:43697|issue:45216", + "issue:43526|issue:45479", + "issue:43720|issue:44898", + "issue:43643|issue:45375", + "issue:44466|issue:45005", + "issue:45200|issue:46463", + "issue:44493|issue:46050", + "issue:41762|issue:44936", + "issue:44361|issue:44442", + "issue:43761|issue:46829", + "issue:43688|issue:43749", + "issue:43452|issue:44843", + "issue:43653|issue:44361", + "issue:43819|issue:44811", + "issue:38175|issue:43295", + "issue:43994|issue:44898", + "issue:45372|issue:46170", + "issue:46832|issue:46932", + "issue:43525|issue:46701", + "issue:43295|issue:45356", + "issue:29159|issue:33357" + ], + "prepared_review_unit_hash": "be6826e1cf4925f85952f7cf147fcf659ef6c323d8286b0a8bbb8bf99d76b0ef", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 36196, + "estimated_input_tokens": 9049, + "estimated_eval_tokens": 18354 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are distinct model-, tokenizer-, or loading-path bugs that only look loosely similar by surface terms. I rejected all soft-edge pairs as non-duplicates. The most broadly representative issue is the multi-model AutoTokenizer regression, but even that does not justify merging the cluster as duplicates.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45920 is the broadest and most central bug report in the set: it describes an AutoTokenizer regression affecting multiple model families, with a detailed reproduction and wider impact than the more model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best single issue to represent the cluster because it is the most general, most widely scoped, and best-documented failure among otherwise unrelated edge-case reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Both are regression-style loading/serialization bugs, but one is specific to Qwen3.5 visual encoder keys while the other is a broad tokenizer output-ID regression across many models; different code paths and failure modes." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both involve RoPE-related config handling, but one is a silent config reload defaulting issue and the other is a type error in JSON-loaded lists during validation; not the same underlying bug." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "An import-time PyTorch missing dependency error is unrelated to a Trainer GPU broadcast NoneType error; different subsystems and no shared code-path bug." + }, + { + "left": "issue:46463", + "right": "issue:46899", + "accept": false, + "reason": "The T5 validation error-message bug and the Gemma4 quantized embedders dtype-cast bug are unrelated; one is messaging/validation, the other is a runtime quantization issue." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU arange placement and flash-attention s_aux=None handling are separate implementation bugs in different modules." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RT-DETR output mismatch and Qwen3.5 incorrect saved checkpoint keys are different regressions affecting different save/load or forward paths." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BEiTImageProcessorFast returning only one reduced label is unrelated to the zero-loss classification problem across sequence-classification models." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet accelerated-loading unpacking and Perceiver interpolation failures are distinct model-specific bugs with no shared concrete code-path." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "One is a generic AutoConfig trust_remote_code field-loss issue; the other is a specific Qwen3_5MoeVisionConfig field being dropped by strict serialization. Similar theme, not the same bug." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is device-dependent lm_head serialization and the other is translation-model tied-weight issues; too broad and not clearly one fixable code-path." + }, + { + "left": "issue:45200", + "right": "issue:46463", + "accept": false, + "reason": "Gemma 4 missing mm_token_type_ids defaults and T5 decoder-input validation error messaging are unrelated bugs." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Unexpected position-id keys and sinq quantization save/load breakage are separate serialization problems." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "DeepSpeed ZeRO-3 loading crash and trainer.evaluate() after train are different training/runtime failures." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Both are tokenizer-related, but one is an AttributeError in MLukeTokenizer on tasks and the other is AutoTokenizer failing to load a specific tokenizer class; not the same underlying defect." + }, + { + "left": "issue:43761", + "right": "issue:46829", + "accept": false, + "reason": "Both concern hidden_states in outputs, but one omits hidden_states entirely while the other misplaces last_hidden_state into hidden_states; different models and output paths." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to FSDP CPU RAM efficient loading." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "gguf_file loading breakage and HF_HUB_OFFLINE model_info calls in mistral regex patch are separate tokenizer/model-loading regressions." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBird special-token decode behavior and MLukeTokenizer task AttributeError are both tokenizer issues but not the same failure mode or code path." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC latent conversion mismatch and Whisper batch_decode skip_special_tokens handling are unrelated model vs processor bugs." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Siglip2 zero probabilities and processor.tokenizer regression are distinct multimodal issues with different symptoms and likely fixes." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Siglip2 nonsensical outputs and Perceiver non-default resolution failure are unrelated model-specific inference bugs." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "A missing mistral_common import for Gemma4 processor loading is unrelated to a Windows ROCm torch.distributed dependency issue." + }, + { + "left": "issue:46832", + "right": "issue:46932", + "accept": false, + "reason": "RT-DETR feature-level mismatch and Gemma4 quantized embedder dtype casting are completely different model paths and fixes." + }, + { + "left": "issue:43525", + "right": "issue:46701", + "accept": false, + "reason": "Llama4Config missing pad_token_id and DiffusionGemmaGenerationOutput pipeline failure are unrelated config vs pipeline issues." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but one is about processor.tokenizer access/images passing and the other is codec handling plus a misleading warning; not one bug." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency is unrelated to a MacOS bus error in a CLIP community model." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 25, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:36010", + "issue:41720", + "issue:42898", + "issue:43452", + "issue:43504", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43611", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43927", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44484", + "issue:44554", + "issue:44568", + "issue:44625", + "issue:44704", + "issue:44734", + "issue:44743", + "issue:44821", + "issue:44843", + "issue:44977", + "issue:44991", + "issue:45440", + "issue:45491", + "issue:45542", + "issue:45701", + "issue:45704", + "issue:46002", + "issue:46050", + "issue:46178", + "issue:46402", + "issue:46491", + "issue:46612", + "issue:46693", + "issue:46821", + "issue:46846" + ], + "soft_pairs": [ + "issue:43653|issue:44568", + "issue:43646|issue:43749", + "issue:43531|issue:43742", + "issue:45542|issue:46178", + "issue:42898|issue:46491", + "issue:43688|issue:45704", + "issue:43756|issue:46402", + "issue:44843|issue:44991", + "issue:43452|issue:43577", + "issue:43653|issue:43927", + "issue:44464|issue:44743", + "issue:43688|issue:45701", + "issue:44479|issue:44625", + "issue:45491|issue:46002", + "issue:44484|issue:44704", + "issue:43688|issue:45440", + "issue:44821|issue:46821", + "issue:33666|issue:36010", + "issue:32090|issue:33666", + "issue:44466|issue:46050", + "issue:44734|issue:46612", + "issue:44977|issue:46693", + "issue:43504|issue:43550", + "issue:44464|issue:44977", + "issue:43504|issue:43720", + "issue:41720|issue:44464", + "issue:43577|issue:44554", + "issue:43611|issue:46846" + ], + "prepared_review_unit_hash": "3b4e64e934e3281bb6c86be9d8ab727f390e4ae0d597e816834176a87b7fcbe1", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35594, + "estimated_input_tokens": 8899, + "estimated_eval_tokens": 18054 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: most items are unrelated Transformers regressions across tokenization, loading, generation, model configs, and backend/device-specific behavior. None of the soft pairs look like true duplicates or the same fixable code-path bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43646 is the broadest representative of the cluster because it concerns a generic Transformers v5 model initialization breakage, but it is not a true duplicate target for most other items.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best single issue to represent the cluster if one must be chosen, since it is high-level and v5-related; however, the surrounding issues are too diverse for strong canonicalization.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer/special-token bugs, but one is BigBird mask-token decoding and the other is mDeBERTa BOS/EOS handling; different failure modes." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization in v5 is unrelated to FSDP CPU RAM efficient loading; different code paths and symptoms." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to a MobileLLM key error while loading pretrained weights." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Both mention backend setup, but one is undefined tf backend after installing tensorboard and the other is a missing BACKENDS_MAPPING entry for tensorflow_text." + }, + { + "left": "issue:42898", + "right": "issue:46491", + "accept": false, + "reason": "One is a v5 clean_up_tokenization_spaces behavior change; the other is a CodeLlama leading-space round-trip regression." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to T5's apex.FusedRMSNorm memory leak." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Smollm3 RoPE-layer selection and Sam3 text-encoder weight loading are different model-specific bugs." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Offline model_info() misuse in _patch_mistral_regex is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf_file loading failures and BLIP2 dtype propagation are distinct issues with different APIs and errors." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBird special-token registration and DiaConfig losing custom token IDs after save/load are different tokenizer/config persistence bugs." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both affect generation with cache, but one is chunked generation under compiled forward and the other is recurrent states resetting in modular_qwen3_5." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "Aux-loss normalization is unrelated to version-dependent tokenization changes." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression on Qwen VL models is unrelated to num_labels propagation in Qwen3.5 configs." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both mention NaNs, but one is Gemma3 sliding-window attention on GPU and the other is classification-head init/checkout behavior." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "A question about save_pretrained shard size is unrelated to AutoProcessor failing to forward kwargs to cached_file." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "MoE aux-loss normalization and DeepseekV3MoE parity with remote implementation are not the same concrete bug." + }, + { + "left": "issue:44821", + "right": "issue:46821", + "accept": false, + "reason": "AutoImageProcessor URL loading and TimesFM window_size AttributeError are unrelated loader/API issues." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and the GenerationMixin import error are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer _gpu_broadcast_one NoneType error is unrelated to Qwen2-VL multi-GPU training." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "lm_head.weight serialization inconsistency and sinq quantized save/load breakage involve different model-state persistence problems." + }, + { + "left": "issue:44734", + "right": "issue:46612", + "accept": false, + "reason": "Wrong tensor indexing in serve KV-cache continuation is different from beam-search cache reorder being skipped for certain models." + }, + { + "left": "issue:44977", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure is distinct from a long-context FA2 performance regression." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Beit legacy-field load failure and Bamba torch.compile+SDPA failure are different model/runtime bugs." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Chunked generation with compiled forward is not the same bug as Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Beit loading with a legacy field and BitNet packed-weight unpacking during accelerate loading are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Qwen3 auto device-mapping cudaErrorAssert on A800 is a different failure than compiled chunked generation inconsistency." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "BLIP2 dtype propagation and MPS attention correctness with mismatched value/query head dims are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:46846", + "accept": false, + "reason": "base_model_prefix loading breakage in v5 and missing linear_attn entries in the TP plan are different config/loading issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 26, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:33666", + "issue:36331", + "issue:39692", + "issue:42796", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43572", + "issue:43606", + "issue:43643", + "issue:43646", + "issue:43688", + "issue:43701", + "issue:43720", + "issue:43782", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44361", + "issue:44451", + "issue:44466", + "issue:44514", + "issue:44554", + "issue:44561", + "issue:44568", + "issue:44610", + "issue:44625", + "issue:44964", + "issue:44977", + "issue:45290", + "issue:45405", + "issue:45584", + "issue:45812", + "issue:45823", + "issue:46050", + "issue:46192", + "issue:46209", + "issue:46365", + "issue:46489", + "issue:46693", + "issue:46710" + ], + "soft_pairs": [ + "issue:43643|issue:44561", + "issue:43782|issue:46693", + "issue:43720|issue:43927", + "issue:33666|issue:36331", + "issue:43278|issue:43688", + "issue:43504|issue:44610", + "issue:43606|issue:43927", + "issue:43262|issue:43295", + "issue:43493|issue:43825", + "issue:43720|issue:44361", + "issue:44964|issue:45405", + "issue:44514|issue:45290", + "issue:44554|issue:46365", + "issue:44977|issue:46209", + "issue:44625|issue:45584", + "issue:43525|issue:43572", + "issue:43452|issue:44451", + "issue:45823|issue:46050", + "issue:28218|issue:33357", + "issue:43994|issue:46192", + "issue:42796|issue:44977", + "issue:43493|issue:43844", + "issue:39692|issue:43295", + "issue:43278|issue:43844", + "issue:44568|issue:46489", + "issue:46209|issue:46710", + "issue:43646|issue:44466", + "issue:44568|issue:45812", + "issue:33357|issue:33666", + "issue:43278|issue:43701" + ], + "prepared_review_unit_hash": "c5cf00e1ee3f38a47d662aadd8de556ac09d65d50c189f78c49a6010c3986639", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36807, + "estimated_input_tokens": 9202, + "estimated_eval_tokens": 18660 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a mix of largely unrelated bug reports. The soft pairs mostly share broad vocabulary (tokenizer/config/regression/processor) but describe different models, code paths, and failure modes, so none should be deduplicated.", + "confidence": 0.79, + "canonical_issue_reason": "If forced to choose one representative issue, issue 45812 is the strongest candidate: it is a concrete tokenizer regression affecting Granite models, has multiple inbound references, and describes a clear code-path bug.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45812 is the best single representative of the cluster\u2019s only recurring theme (tokenizer regressions), with the clearest impact and most actionable report among the listed issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both involve v5/trust_remote_code breakage, but one is missing fields from AutoConfig and the other is removal of is_torch_fx_available; different bugs and fixes." + }, + { + "left": "issue:43782", + "right": "issue:46693", + "accept": false, + "reason": "Qwen3VL from_pretrained weight_only failure and FA2 long-context performance regression are unrelated code paths." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet packed-weight loading vs DiaConfig token-ID loss on save/load are distinct failures." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Multi-GPU Qwen2-VL training and a CustomTrainer compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Embedding dtype drift during evaluate is not the same as auxiliary-loss normalization in MoE models." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Legacy field loading for Beit semantic segmentation and OmDet processor image-size mismatch are different bugs." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch and DiaConfig custom token-ID corruption are different failure modes." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio processor sampling-rate defaulting and processor.tokenizer/image handling regression do not share the same underlying defect." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and a pipeline error message regression are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Packed-weight loading in BitNet and MLukeTokenizer AttributeError are distinct issues." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal model loading failure and an unreleased PEFT version bump are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both hit apply_chat_template, but one is batched padding=False on Qwen2.5-VL and the other is assistant tool-call handling; different edge cases." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness and importing Trainer failing due to datasets are completely unrelated." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 flash-attention generation problems and FP8 inference on B200 are different bugs." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config and missing pad_token_idx in StableLmConfig are similar naming, but different configs and different breakages." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file breaking tokenizer/model loading and a ScandiBERT load failure are not the same defect." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Gemma4 device-map mismatch and sinq save/load quantization breakage are unrelated." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing after added tokens and a MacOS bus error in CLIP are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and return_tensors being ignored when text_kwargs is passed are different issues." + }, + { + "left": "issue:42796", + "right": "issue:44977", + "accept": false, + "reason": "Ministral BPE marker handling and Qwen3.5 flash-attention generation are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "A model discrepancy and a DeepSpeed gradient regression do not share the same code-path problem." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Documentation/example errors for SigLIP2 are not the same as the processor.tokenizer/images regression." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype change and ZeRO-3 gradient growth are unrelated training behaviors." + }, + { + "left": "issue:44568", + "right": "issue:46489", + "accept": false, + "reason": "Both are tokenizer bugs, but they affect different models and mechanisms (special-token insertion vs DeepSeek-Coder BPE regression)." + }, + { + "left": "issue:46209", + "right": "issue:46710", + "accept": false, + "reason": "FP8 inference failure on B200 and an accuracy regression after upgrade are different problems." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage in v5 and inconsistent lm_head serialization are distinct issues." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both involve tokenization, but one is add_special_tokens behavior for mdeberta and the other is wrong token IDs for Granite; not the same bug." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "MacOS bus error on CLIP and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Embedding dtype drift in evaluation and resume_from_checkpoint key mismatch are different defects." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 27, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:43257", + "issue:43299", + "issue:43334", + "issue:43504", + "issue:43572", + "issue:43575", + "issue:43606", + "issue:43653", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43937", + "issue:43994", + "issue:44112", + "issue:44291", + "issue:44361", + "issue:44464", + "issue:44488", + "issue:44493", + "issue:44589", + "issue:44792", + "issue:44843", + "issue:44849", + "issue:44869", + "issue:44964", + "issue:45072", + "issue:45200", + "issue:45229", + "issue:45245", + "issue:45357", + "issue:45405", + "issue:45520", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46050", + "issue:46209", + "issue:46285", + "issue:46399", + "issue:46424", + "issue:46701", + "issue:46710", + "issue:46752", + "issue:46821", + "issue:46832", + "issue:46846" + ], + "soft_pairs": [ + "issue:44488|issue:44843", + "issue:44849|issue:46209", + "issue:44792|issue:45405", + "issue:43927|issue:44361", + "issue:43575|issue:45405", + "issue:43504|issue:43606", + "issue:44849|issue:45357", + "issue:43824|issue:44589", + "issue:43257|issue:43299", + "issue:45245|issue:46710", + "issue:44493|issue:46285", + "issue:46399|issue:46846", + "issue:46821|issue:46832", + "issue:43606|issue:44112", + "issue:43334|issue:45759", + "issue:43334|issue:43572", + "issue:43504|issue:43653", + "issue:44589|issue:45520", + "issue:45072|issue:46832", + "issue:43572|issue:44291", + "issue:44464|issue:46424", + "issue:43937|issue:46701", + "issue:43844|issue:44964", + "issue:43873|issue:46050", + "issue:45200|issue:45229", + "issue:44869|issue:46752", + "issue:45229|issue:45823", + "issue:43994|issue:44493", + "issue:45588|issue:46050" + ], + "prepared_review_unit_hash": "fa7b28f5672b2c5a5090712f76ed79be6f5569b74aa09bd2c7044fe2cf5e8586", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35959, + "estimated_input_tokens": 8990, + "estimated_eval_tokens": 18236 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous. The clearest subcluster is Qwen3 MoE loading regressions, with 43299 as the broadest representative. Qwen multimodal/config-loading reports are related by theme but not strong enough to treat as the same bug, especially 43334 vs 45759.", + "confidence": 0.31, + "canonical_issue_reason": "issue:43299 is the broadest early Qwen3/Qwen3VL MoE loading regression in the packet, so it best represents the small loading-failure subcluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43299 is the best single anchor for the cluster\u2019s only clearly related issue group; the rest of the packet is too heterogeneous to justify a broader canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Both are loading-related, but one is a specific model-load failure and the other is an offline-mode tokenizer patch regression; different code paths." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Different model families and failure modes: Qwen3.5 hidden-state behavior vs FP8 inference on B200." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus generation test failure vs a PEFT version packaging issue; not the same bug." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Distinct tokenizer/config problems with different symptoms and affected models." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Tensor-parallel OOM during loading is unrelated to the PEFT version bump issue." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Legacy-field model loading and CPU offload device mismatch are different bugs." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Same model family, but hidden-states behavior and save_pretrained key regression are different concrete issues." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Import error for a Qwen class and Float8 storage lookup failure are unrelated compatibility problems." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": true, + "reason": "Both are Qwen3/Qwen3-VL MoE loading regressions on the dev/5.0 line; a shared fix in the loading/conversion path could plausibly cover both." + }, + { + "left": "issue:45245", + "right": "issue:46710", + "accept": false, + "reason": "Category-count runtime error and an accuracy regression are unrelated." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Unexpected position-id keys are unrelated to the CLIP text-model structural change." + }, + { + "left": "issue:46399", + "right": "issue:46846", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is a kernel registration error and the other is a tensor-parallel plan/OOM issue." + }, + { + "left": "issue:46821", + "right": "issue:46832", + "accept": false, + "reason": "TimesFM window_size AttributeError and RT-DETR feature-level mismatch are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "The CI test failure is not the same bug as the Bark CPU-offload device mismatch." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Related Qwen multimodal/config-loading area, but the packet only shows distinct titles and symptoms; not enough evidence they are the same underlying bug." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Both are config-field regressions after 5.0, but they affect different models and different missing/changed fields." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy-field model loading and tokenizer special-token registration are unrelated." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Float8 storage lookup and flash_attn import on Python 3.13 are different compatibility issues." + }, + { + "left": "issue:45072", + "right": "issue:46832", + "accept": false, + "reason": "Dtype mismatch in inference and RT-DETR feature-level indexing/type error are different subsystems." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Missing config attribute and unexpected init argument are related only at a high level, not the same fix." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is unrelated to cache max length handling." + }, + { + "left": "issue:43937", + "right": "issue:46701", + "accept": false, + "reason": "GenerationConfig validation and DiffusionGemma pipeline output issues are unrelated." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient instability and a multimodal model load failure are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantization/offloading and sinq save/load breakage are different code paths." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Text-only token defaults and multi-GPU CUDA OOM are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:46752", + "accept": false, + "reason": "Whisper timestamp decoding crash and empty chat-template IndexError are different edge cases." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma4 inference issues, but one is OOM and the other is a device mismatch; not the same bug." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 incorrect outputs and unexpected position-id keys are unrelated." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention sink-less-model crash is unrelated to sinq serialization/deserialization breakage." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Mostly grounded, but one soft-edge acceptance is too strong for the evidence available. In particular, issue:43334 vs issue:45759 looks related at the Qwen multimodal/config level, but the packet only shows distinct titles ('missing pad_token_id' vs 'does not unwrap text_config') and does not justify treating them as likely the same underlying bug. That makes the summary's 'plausible duplicate subclusters' claim a bit overstated. The other soft-edge rejections look conservative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 28, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:29159", + "issue:29870", + "issue:29942", + "issue:30333", + "issue:32090", + "issue:34567", + "issue:37428", + "issue:38175", + "issue:41720", + "issue:42490", + "issue:43065", + "issue:43295", + "issue:43404", + "issue:43450", + "issue:43493", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43696", + "issue:43701", + "issue:43844", + "issue:43873", + "issue:43957", + "issue:44186", + "issue:44423", + "issue:44554", + "issue:44610", + "issue:44734", + "issue:44977", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45588", + "issue:45759", + "issue:45823", + "issue:46121", + "issue:46178", + "issue:46192", + "issue:46369", + "issue:46701", + "issue:46710", + "issue:46821" + ], + "soft_pairs": [ + "issue:29942|issue:30333", + "issue:42490|issue:43696", + "issue:29159|issue:32090", + "issue:28218|issue:34567", + "issue:44977|issue:46701", + "issue:45084|issue:45520", + "issue:43065|issue:46369", + "issue:44554|issue:45446", + "issue:15354|issue:38175", + "issue:44423|issue:44734", + "issue:41720|issue:46710", + "issue:43493|issue:43701", + "issue:45406|issue:46121", + "issue:43404|issue:46710", + "issue:45072|issue:45198", + "issue:43295|issue:43450", + "issue:45588|issue:45759", + "issue:37428|issue:43957", + "issue:43540|issue:45200", + "issue:43295|issue:43873", + "issue:44186|issue:46192", + "issue:45823|issue:46192", + "issue:43295|issue:44610", + "issue:43572|issue:46821", + "issue:43550|issue:43844", + "issue:45200|issue:45823", + "issue:29870|issue:29942", + "issue:45520|issue:46178" + ], + "prepared_review_unit_hash": "92e174fde8eba83be6fc6d04e23c713543aaea116878e240f3d40d82434d4461", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36238, + "estimated_input_tokens": 9060, + "estimated_eval_tokens": 18376 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly false positives: they share broad subsystems like tokenizers, serving, or multimodal models, but not the same concrete bug. No PRs are present.", + "confidence": 0.86, + "canonical_issue_reason": "issue:29159 is the broadest, most reusable tokenizer-behavior report among the candidates and the best umbrella-style issue, even though it does not match the soft pairs as duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:29159 is the most general and actionable issue to anchor this set; it is well-scoped and likely the most useful canonical representative if these were being grouped loosely by tokenizer behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Different bugs: Flash Attention 2 test failures vs MLflow reporting after training failure." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Unrelated: local-path import warning vs CUDA OOM on GPT-oss-20b." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency is unrelated to a Trainer NoneType GPU broadcast error." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer added-space behavior and TrainerState token counting are different code paths and symptoms." + }, + { + "left": "issue:44977", + "right": "issue:46701", + "accept": false, + "reason": "Both involve generation, but different model families and failure modes; not the same bug." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Non-template-node compile error is unrelated to flash_attn import-key errors." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "Same model family, but dummy Conv2d construction and memory overflow are not the same defect." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "Attention correctness on MPS is unrelated to a PyTorch version check bug in flex_attention imports." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure is unrelated to SigLIP2 zero probabilities." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are serve crashes, but they hit different paths: multimodal tensor conversion vs KV-cache continuation indexing." + }, + { + "left": "issue:41720", + "right": "issue:46710", + "accept": false, + "reason": "Different models and problems: cudaErrorAssert with auto device mapping vs an accuracy regression after upgrade." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 parity discrepancy and resume_from_checkpoint key mismatch are unrelated." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor attribute crash and rope-param JSON list handling are different bugs." + }, + { + "left": "issue:43404", + "right": "issue:46710", + "accept": false, + "reason": "Weight tying in Mistral3 is unrelated to DeepSeek accuracy regression." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "bfloat16 dtype mismatch tests are unrelated to save_pretrained/tokenization failures." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor-tokenizer regression and batched video shape bug are different issues." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Flash-attention sink handling and Qwen multimodal config unwrapping are separate code paths." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Import name change in flash_attention_utils is unrelated to meta-device model loading failures." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video processing ValueError and Gemma4 mm_token_type_ids defaulting are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor API regression is unrelated to quantization offloading behavior." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "Both mention tokenizers, but NER/padding crashes and ignored return_tensors are distinct bugs." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Gemma4 device mismatch is unrelated to return_tensors being ignored." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor/tokenizer regression and OmDet-Turbo image size mismatch are not the same bug." + }, + { + "left": "issue:43572", + "right": "issue:46821", + "accept": false, + "reason": "Different config/attribute issues in unrelated models." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile SDPA failure and ZeRO-3 gradient growth are unrelated training/runtime bugs." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but text-only fine-tuning defaults and forward device mismatch are different failures." + }, + { + "left": "issue:29870", + "right": "issue:29942", + "accept": false, + "reason": "Bart generate masking error is unrelated to Flash Attention 2 test failures." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "flash_attn import error on Python 3.13 and BACKENDS_MAPPING tensorflow_text error are different backend-registration issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 29, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:37428", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43450", + "issue:43452", + "issue:43493", + "issue:43525", + "issue:43572", + "issue:43575", + "issue:43653", + "issue:43819", + "issue:43824", + "issue:43856", + "issue:43950", + "issue:43975", + "issue:44112", + "issue:44442", + "issue:44466", + "issue:44521", + "issue:44554", + "issue:44737", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44987", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45356", + "issue:45381", + "issue:45412", + "issue:45563", + "issue:46192", + "issue:46209", + "issue:46285", + "issue:46512", + "issue:46894", + "issue:46899", + "issue:46932" + ], + "soft_pairs": [ + "issue:43450|issue:43575", + "issue:43572|issue:44987", + "issue:44554|issue:45356", + "issue:42490|issue:43493", + "issue:37428|issue:43824", + "issue:39692|issue:46192", + "issue:43232|issue:44849", + "issue:44466|issue:45127", + "issue:43295|issue:43493", + "issue:44521|issue:45290", + "issue:45381|issue:46209", + "issue:43653|issue:44112", + "issue:43278|issue:43856", + "issue:28218|issue:35141", + "issue:44737|issue:46894", + "issue:43824|issue:43975", + "issue:43950|issue:46285", + "issue:42915|issue:43824", + "issue:43452|issue:46512", + "issue:43262|issue:43450", + "issue:44442|issue:45198", + "issue:45200|issue:46932", + "issue:45200|issue:46899", + "issue:43334|issue:43525", + "issue:44792|issue:45563", + "issue:44898|issue:45412", + "issue:43819|issue:44112" + ], + "prepared_review_unit_hash": "f28cf8134459d443aae546a1083c702edc67f1492e761e368abf5ce4b86a3645", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36188, + "estimated_input_tokens": 9047, + "estimated_eval_tokens": 18350 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous; almost all soft edges are false positives. The only plausible duplicate pair is the two missing-`pad_token_id` config-loading regressions.", + "confidence": 0.84, + "canonical_issue_reason": "If a canonical issue must be chosen, issue:43334 is the best anchor because it captures the shared `pad_token_id`-missing regression that issue:43525 appears to duplicate; the rest of the items are unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43334 is the strongest representative of the only real duplicate subcluster (shared `pad_token_id` AttributeError during model loading).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Different problems: batched video shape vs tensor-parallel OOM." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading failures, but for different models and different root causes." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness bug vs tokenizer regression; unrelated code paths." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Local-path slash warning is unrelated to SigLIP2 implementation discrepancy." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Different ImportErrors involving different symbols and code paths." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "Documentation-example issues vs `return_tensors` handling; not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwargs update bug vs Qwen3.5 hidden-states bug; different failures." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both involve tied weights, but one is serialization and the other is LoRA merge collapse." + }, + { + "left": "issue:43295", + "right": "issue:43493", + "accept": false, + "reason": "Processor/tokenizer regression vs SigLIP2 implementation mismatch." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both touch `apply_chat_template`, but one is zeroed assistant masks and the other is a crash on tool-call messages." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Video `vision_position_ids` bug vs FP8 inference failure; unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token registration issue vs GraniteSpeech CI device override test." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "dtype drift in evaluation vs MoE training memory usage." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing bug vs embedding reinitialization during resize; different bugs." + }, + { + "left": "issue:44737", + "right": "issue:46894", + "accept": false, + "reason": "XLNet CPU tensor allocation issue vs ASR double-shift training-loss bug." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Missing export ImportError vs incorrect detokenization; unrelated." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption vs CLIP text-model structure inconsistency." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "FP8-config model failure vs missing Qwen2.5-VL import; different code paths." + }, + { + "left": "issue:43452", + "right": "issue:46512", + "accept": false, + "reason": "`gguf_file` loading breakage vs stale `config.dtype` after `from_config`." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sample-rate default vs batched video output shape; different processors." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Tokenizer auto-loading failure vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:45200", + "right": "issue:46932", + "accept": false, + "reason": "Missing multimodal token-type defaults vs quantized embedder dtype casting; not the same bug." + }, + { + "left": "issue:45200", + "right": "issue:46899", + "accept": false, + "reason": "Text-only `mm_token_type_ids` defaulting vs unsafe dtype cast in quantized embedders." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": true, + "reason": "Both report the same `pad_token_id`-missing AttributeError in model/config loading, likely fixable with one shared fallback or config change." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure vs stale paged-generate warning." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver interpolation failure vs RT-DETR memory release issue." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC forward/from_latents mismatch vs GraniteSpeech CI device override regression." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is conservative and grounded in the packet. The false soft-edge verdicts mostly stay at the title/symptom level and do not overclaim shared root causes. The only accepted pair, issue:43334 with issue:43525, is supported by the shared `pad_token_id` AttributeError symptom in the titles, and the rest of the set does look heterogeneous. No mergeability concern here since these are issue pairs, not PR pairs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 30, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28282", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:37428", + "issue:38175", + "issue:41720", + "issue:43335", + "issue:43388", + "issue:43425", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43827", + "issue:43901", + "issue:44112", + "issue:44206", + "issue:44448", + "issue:44466", + "issue:44479", + "issue:44561", + "issue:44610", + "issue:44743", + "issue:44855", + "issue:44857", + "issue:44995", + "issue:45083", + "issue:45200", + "issue:45325", + "issue:45412", + "issue:45520", + "issue:45584", + "issue:45593", + "issue:45759", + "issue:46170", + "issue:46192", + "issue:46209", + "issue:46234", + "issue:46285", + "issue:46424", + "issue:46512", + "issue:46821", + "issue:46858" + ], + "soft_pairs": [ + "issue:44479|issue:46192", + "issue:45200|issue:45325", + "issue:41720|issue:46209", + "issue:43827|issue:43901", + "issue:43335|issue:43550", + "issue:44561|issue:46234", + "issue:43819|issue:45083", + "issue:43388|issue:43526", + "issue:45759|issue:46285", + "issue:44857|issue:45593", + "issue:44610|issue:44857", + "issue:46424|issue:46858", + "issue:44995|issue:46858", + "issue:43606|issue:45083", + "issue:46512|issue:46821", + "issue:44206|issue:45584", + "issue:38175|issue:45412", + "issue:43335|issue:43653", + "issue:29942|issue:32090", + "issue:43550|issue:44112", + "issue:28282|issue:33357", + "issue:37428|issue:43425", + "issue:43611|issue:44466", + "issue:44448|issue:44743", + "issue:43335|issue:43504", + "issue:45520|issue:46170", + "issue:44855|issue:45083" + ], + "prepared_review_unit_hash": "67e3afaa342c1607671de2937cbc2f09d9586924fcc16107824af67e34353c05", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35640, + "estimated_input_tokens": 8910, + "estimated_eval_tokens": 18076 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated bug reports across different models, APIs, and environments. I did not find any soft pair that clearly describes the same underlying bug/change, so no duplicates were accepted and no canonical artifact stands out.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44479", + "right": "issue:46192", + "accept": false, + "reason": "Both concern multimodal input handling, but one is a video regression in several Qwen VL models while the other is a return_tensors/text_kwargs precedence bug; different code paths." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "One is about missing mm_token_type_ids defaults for text-only fine-tuning; the other is a rope/position_ids scaling bug in Qwen2.5-VL. Different behavior and fix surface." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Both are device/hardware failures, but one is a cudaErrorAssert on A800 with Qwen3 auto device mapping and the other is FP8 inference failure on B200. Not the same bug." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues, but they cover different APIs and deprecations: pipeline() removal versus return_all_scores wording." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Different model families and failure modes: SwitchTransformers config creates an unwanted sparse layer, while Bamba fails under torch.compile with SDPA." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Both mention missing import_utils helpers, but they are different symbols affecting different models and dependencies." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC.from_latents vs qwen3_omni_moe feature-length helper are unrelated model internals and bugs." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both involve labels, but one is gather_for_metrics truncating the last batch and the other is BeitImageProcessorFast.reduce_labels returning only one label. Different root causes." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "One is an AutoModelForCausalLM config-unwrapping issue for Qwen multimodal configs; the other is a CLIP text model structure inconsistency." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Different detection/training bugs: float16 AMP crash in LwDetrImageLoss versus missing auxiliary losses in D-FINE when denoising is off." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Different models and symptoms: OmDet-Turbo processor output size mismatch versus LwDetrImageLoss AMP crash." + }, + { + "left": "issue:46424", + "right": "issue:46858", + "accept": false, + "reason": "Both are cache-related, but one is max_cache_length only applying to Quantized Cache while the other is get_seq_length returning a device tensor that breaks slice arithmetic under torch.compile." + }, + { + "left": "issue:44995", + "right": "issue:46858", + "accept": false, + "reason": "Both mention cache behavior, but one is a stale indexer cache causing second-forward crashes and the other is a device-tensor seq_length arithmetic bug." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU offload device mismatch in bark-small is unrelated to qwen3_omni_moe feature-extraction length behavior." + }, + { + "left": "issue:46512", + "right": "issue:46821", + "accept": false, + "reason": "Different API regressions: stale config.dtype after from_config(dtype=...) versus AttributeError for TimesFM 2.5 window_size." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "LasrFeatureExtractor center-arg crash and Whisper empty-transcription failure are distinct preprocessing/generation bugs." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Unrelated symptoms and components: zero probabilities in SigLIP2 versus RT-DETR memory not being released." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Different areas: SwitchTransformers sparse-layer creation versus BigBirdTokenizer special-token registration and decode output." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and _gpu_broadcast_one NoneType errors are unrelated test/runtime issues." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba SDPA/torch.compile failure and GraniteSpeech stale device-override CI failure do not share the same code path." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "ImportError about missing PyTorch and a MacOS bus error in CLIP model loading are different failure modes." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Importing a missing flash-attention helper and a Torch 2.10 compatibility report are too broad and not the same bug." + }, + { + "left": "issue:43611", + "right": "issue:44466", + "accept": false, + "reason": "base_model_prefix loading breakage and lm_head.weight serialization inconsistency are separate v5 regressions." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output differences and recurrent state reset in modular_qwen3_5 are unrelated model behavior changes." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Different bugs in different models: sparse layer creation in SwitchTransformers versus legacy field loading in BeitForSemanticSegmentation." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Python 3.13 flash_attn import-key issue and Windows ROCm torch.distributed import/type errors are different platform-specific problems." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "DebertaV2Model import parsing on Python 3.13 is unrelated to qwen3_omni_moe feature-length helper behavior." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 31, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:26413", + "issue:28218", + "issue:29127", + "issue:31501", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:39692", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43526", + "issue:43575", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43819", + "issue:43825", + "issue:43854", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44448", + "issue:44493", + "issue:44514", + "issue:44521", + "issue:44617", + "issue:44779", + "issue:44857", + "issue:45072", + "issue:45083", + "issue:45127", + "issue:45245", + "issue:45290", + "issue:45310", + "issue:45459", + "issue:45478", + "issue:45588", + "issue:45593", + "issue:45823", + "issue:45859", + "issue:46234", + "issue:46326", + "issue:46431", + "issue:46566", + "issue:46697" + ], + "soft_pairs": [ + "issue:43278|issue:43825", + "issue:43335|issue:45083", + "issue:43335|issue:45593", + "issue:43335|issue:44112", + "issue:29127|issue:32090", + "issue:43720|issue:44112", + "issue:44514|issue:44521", + "issue:45290|issue:46326", + "issue:45588|issue:45859", + "issue:42490|issue:43854", + "issue:37428|issue:44351", + "issue:34689|issue:35141", + "issue:45127|issue:45245", + "issue:39692|issue:43526", + "issue:43819|issue:44493", + "issue:46431|issue:46566", + "issue:26413|issue:31501", + "issue:44448|issue:44779", + "issue:45310|issue:45478", + "issue:43606|issue:44857", + "issue:28218|issue:32090", + "issue:45459|issue:46234", + "issue:39692|issue:44493", + "issue:43701|issue:45083", + "issue:44617|issue:46697", + "issue:44206|issue:45823", + "issue:32090|issue:33357", + "issue:42490|issue:43575", + "issue:44857|issue:45072" + ], + "prepared_review_unit_hash": "7d0a761909637c5fe47feea7d6f433b1a00896d3a3888c0f34d3a9dcfa7ec8af", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36496, + "estimated_input_tokens": 9124, + "estimated_eval_tokens": 18504 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated bug reports; only one clear duplicate pair stands out. The strongest match is the duplicated Qwen3.5 MoE from_pretrained regression (45310/45478).", + "confidence": 0.93, + "canonical_issue_reason": "issue:45310 is the best canonical issue because it has an exact duplicate partner (issue:45478) with the same title and the same underlying Qwen3.5 MoE from_pretrained failure, and it appears earlier in time.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the most representative issue in the cluster: it captures the clearest duplicate theme with the strongest textual match and enough discussion to serve as the anchor report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Different problems: one is a dtype change during train/eval, the other is a pipeline error message about translation support." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Different code paths: SwitchTransformers sparse-layer construction vs qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "Unrelated issues: sparse layer creation in SwitchTransformers vs missing auxiliary losses in D-FINE." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "No concrete overlap: SwitchTransformers config bug vs a GraniteSpeech CI device override test failure." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "Different symptoms and subsystems: LayoutLMv3 error messaging vs a Trainer/_gpu_broadcast_one NoneType type error." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Different bugs in different components; no shared failing code path." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both involve apply_chat_template, but one is a batched padding=False crash and the other is incorrect all-zero assistant masks for multimodal inputs; not the same concrete bug." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both touch chat templating, but one is tool-call handling with empty assistant content and the other is continue_final_message behavior for mistral_common tokenizers." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Different failure modes and code paths: flash_attention sink-less s_aux handling vs Qwen3_5MoeTextRotaryEmbedding CPU-offload compatibility." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Not the same bug: local-path-with-slash import warning vs a unit-test model loading failure for GLM-4.7-Flash." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols in different modules." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Different regressions: Llama 3.2 Vision model loading vs post_init reinitializing output embeddings after resize when tie_word_embedding is False." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "Completely different bugs: LoRA/extended-vocab collapse vs category-count overflow." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "Different model areas and failures: SigLIP2 docs/model-processor issues vs BEiT image processor label reduction bug." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "Different concrete problems: DAC from_latents mismatch vs unexpected position-id key warnings across many models." + }, + { + "left": "issue:46431", + "right": "issue:46566", + "accept": false, + "reason": "No shared code path: missing chat template on a processor vs DiffusionGemma staying on meta with device_map='auto'." + }, + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "One is a resume_from_checkpoint epoch-iterator failure; the other is a FineWeb training startup issue with different root causes." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 regressions, but one is Pegasus output drift and the other is Deepseek tokenizer correctness; different artifacts and fixes." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Exact duplicate reports of the same Qwen3.5 MoE from_pretrained error, with matching title and underlying failure." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "Different models and failures: CPU offload device mismatch in bark-small vs AMP CUDA crash in LwDetrImageLoss." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing bug vs Trainer NoneType broadcast error; unrelated." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Different issues: tokenizer error masking when protobuf is missing vs a missing availability helper in import_utils." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "Different subjects: SigLIP2 documentation/model setup issues vs unexpected position-id keys across many models." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Both are bug reports, but one is resume_from_checkpoint key mismatch and the other is a qwen3_omni_moe feature-length helper issue." + }, + { + "left": "issue:44617", + "right": "issue:46697", + "accept": false, + "reason": "Different performance/correctness problems: CUDA OOM in Sam3Video vs over-padding in Swin2SR image preprocessing." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Different issues in different models/components: unsupported center arg in LasrFeatureExtractor vs Gemma4 PLE device mismatch." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Different failure modes and platforms: NoneType broadcast type error vs MacOS bus error with a CLIP model." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Different bugs: path-ending-with-slash import warning vs tensor-parallel load OOM." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Both are dtype-related, but they affect different models and code paths; not the same concrete bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet. The only accepted duplicate pair, issue:45310/issue:45478, is strongly supported by the identical title and high textual similarity, and the other soft-edge rejections are phrased conservatively without overclaiming cross-issue sameness. The canonical choice is also reasonable given the earlier timestamp." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 32, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:25251", + "issue:27957", + "issue:29870", + "issue:29942", + "issue:32090", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:40444", + "issue:42175", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43526", + "issue:43550", + "issue:43873", + "issue:43927", + "issue:43976", + "issue:44360", + "issue:44448", + "issue:44466", + "issue:44493", + "issue:44561", + "issue:44568", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44871", + "issue:44977", + "issue:45083", + "issue:45229", + "issue:45290", + "issue:45381", + "issue:45440", + "issue:45491", + "issue:45588", + "issue:45698", + "issue:45920", + "issue:46192", + "issue:46399", + "issue:46421", + "issue:46655", + "issue:46858" + ], + "soft_pairs": [ + "issue:43550|issue:45083", + "issue:44360|issue:44561", + "issue:29942|issue:35141", + "issue:27957|issue:34689", + "issue:38175|issue:44493", + "issue:43278|issue:43329", + "issue:29870|issue:32090", + "issue:45920|issue:46192", + "issue:44448|issue:44977", + "issue:45290|issue:45381", + "issue:44977|issue:45229", + "issue:10000|issue:15354", + "issue:43927|issue:45698", + "issue:44855|issue:46399", + "issue:40444|issue:43526", + "issue:43278|issue:43381", + "issue:29870|issue:33290", + "issue:45290|issue:46655", + "issue:44857|issue:45588", + "issue:32090|issue:33290", + "issue:37428|issue:42175", + "issue:44561|issue:44779", + "issue:46421|issue:46858", + "issue:43873|issue:45229", + "issue:44448|issue:44466", + "issue:42490|issue:43873", + "issue:43976|issue:45440", + "issue:25251|issue:32090", + "issue:44568|issue:45491", + "issue:44871|issue:44977" + ], + "prepared_review_unit_hash": "79d984eb16b4a0dc11abbe077f7455b3e2bbb8f8d46a9ef31e9a6a9232395914", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37064, + "estimated_input_tokens": 9266, + "estimated_eval_tokens": 18788 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated bug reports across different subsystems and models. I did not find any soft pair that clearly describes the same underlying bug/change, so no duplicate edges are accepted. The broad tokenizer regression in issue 45920 is the best representative issue in this set.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45920 is the broadest and most central issue here: it describes a multi-model tokenizer regression in v5, matches several downstream symptoms, and has the strongest cluster-wide relevance.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best issue candidate because it captures a wider user-facing regression rather than a model-specific or narrow edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Different failures: torch.compile/SDPA crash vs a qwen3_omni_moe feature-length helper bug." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "Unrelated code paths: DSA indexer/ReLU discussion vs removal of is_torch_fx_available breaking trust_remote_code." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are not the same as output-embedding reinitialization during resize/post_init." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "One is XLMRoberta Flash Attention 2 support; the other is a Llama 3.2 vision model loading regression." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Unexpected zero probabilities and unexpected position-id keys are different symptoms with no clear shared bug." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "BF16-to-FP32 eval dtype drift is unrelated to undefined variables in the multimodal video token counter." + }, + { + "left": "issue:29870", + "right": "issue:32090", + "accept": false, + "reason": "Generate-time decoder_attention_mask bug and Trainer GPU broadcast NoneType error are separate code paths." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Tokenizer ID regression across models is not the same as return_tensors being ignored when text_kwargs is passed." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus v4/v5 output mismatch and Qwen3.5 flash-attention generation failure are different model-specific regressions." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat template crash on tool-call assistant messages is unrelated to video vision_position_ids being wrong." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Flash-attention generation bug and multi-GPU CUDA OOM are different problems." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "Different model/export failures: DistilBert issue vs torch.jit.script generator-expression export limitation." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load, but one loses custom token IDs in config while the other loads the wrong custom module." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Python 3.13 IndentationError in DebertaV2Model is unrelated to a missing decorator in Qwen3.5 kernel registration." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Multi-image iterable finetuning failure and BeitImageProcessorFast reduce_labels bug are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Eval dtype change and gradient-checkpointing-in-eval limitation are distinct issues." + }, + { + "left": "issue:29870", + "right": "issue:33290", + "accept": false, + "reason": "Generate() decoder attention mask bug is unrelated to DeepSpeed Adafactor OOM." + }, + { + "left": "issue:45290", + "right": "issue:46655", + "accept": false, + "reason": "Chat-template tool-call crash and video padding miscalculation are different features and failures." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "AMP/CUDA crash in LwDetrImageLoss is unrelated to flash_attention.py failing on s_aux=None." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer NoneType broadcast error and DeepSpeed optimizer OOM are different problems." + }, + { + "left": "issue:37428", + "right": "issue:42175", + "accept": false, + "reason": "Missing flash_attention_utils import and TensorFlow not included in the torch extra are unrelated." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Trust_remote_code breakage from removed helper is not the same as DeepSeek tokenizer regression." + }, + { + "left": "issue:46421", + "right": "issue:46858", + "accept": false, + "reason": "Lazy cache initialization causing recompilation is different from get_seq_length returning a device tensor in static cache arithmetic." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization offloading behavior and Gemma4 multi-GPU OOM are not the same bug." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python-version compatibility regression is unrelated to DeepSeekV3MoE diverging from the remote implementation." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Model output differences and lm_head.weight serialization inconsistency are separate issues." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Local-path import warning has no relation to quantization offloading behavior." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "top_k nesting in pipeline output is unrelated to the Trainer GPU broadcast NoneType error." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer BOS/EOS omission and Gemma3 NaN embeddings are different bugs." + }, + { + "left": "issue:43329", + "right": "issue:45381", + "accept": false, + "reason": "Undefined video helper variables and incorrect vision_position_ids are related to video processing broadly, but not clearly the same underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 33, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:29127", + "issue:29870", + "issue:29942", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:39692", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43278", + "issue:43381", + "issue:43412", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44361", + "issue:44442", + "issue:44479", + "issue:44743", + "issue:44977", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45593", + "issue:45712", + "issue:45803", + "issue:45823", + "issue:46121", + "issue:46693" + ], + "soft_pairs": [ + "issue:29127|issue:34567", + "issue:45588|issue:45823", + "issue:41720|issue:45229", + "issue:43381|issue:43701", + "issue:33357|issue:39692", + "issue:43278|issue:43575", + "issue:44977|issue:45563", + "issue:45198|issue:45803", + "issue:44186|issue:45198", + "issue:44361|issue:45198", + "issue:43611|issue:46121", + "issue:43653|issue:44743", + "issue:29870|issue:34689", + "issue:45372|issue:45588", + "issue:43756|issue:45440", + "issue:42915|issue:42994", + "issue:44206|issue:44479", + "issue:29942|issue:30064", + "issue:43653|issue:44442", + "issue:45459|issue:45712", + "issue:44265|issue:45198", + "issue:45491|issue:46693", + "issue:45372|issue:45406", + "issue:43065|issue:43994", + "issue:43412|issue:45588", + "issue:29127|issue:33357", + "issue:45245|issue:45563", + "issue:43065|issue:45412", + "issue:43575|issue:43701", + "issue:45440|issue:45593", + "issue:42490|issue:43381", + "issue:45372|issue:45459", + "issue:29127|issue:35141" + ], + "prepared_review_unit_hash": "7fc93973aaaf53341920a088109cfffa4106e9fd65a3bb3fb66e8e0dcc10a8fb", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37674, + "estimated_input_tokens": 9419, + "estimated_eval_tokens": 19094 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: most items are unrelated product/model-specific bugs, and none of the soft pairs look like true duplicates or the same fixable change. No soft edges should be merged.", + "confidence": 0.59, + "canonical_issue_reason": "If a single anchor is required, 46693 is the broadest and most current regression issue in the set (FA2/long-context generation), and it is the only item with any meaningful thematic proximity to another candidate. The cluster overall is not a true duplicate cluster, though.", + "canonical_pr_reason": null, + "best_issue_reason": "46693 is the least-specific, most central issue among otherwise unrelated reports, so it is the best representative anchor for triage. It still should not be treated as a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "Different models and failures: LayoutLMv3 error-message clarity vs TrainerState token-count update bug." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Different code paths: flash attention None handling vs Gemma4 PLE device placement mismatch." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Both are OOM/compute failures, but on different models and causes; not the same bug." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Unrelated failures: eval-mode gradient checkpointing restriction vs checkpoint key mismatch." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error on a community CLIP model vs SigLIP2 documentation/model-processor mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "Different issues: dtype regression in eval vs tensor-parallel load OOM." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation bug vs a stale warning for num_return_sequences; not the same defect." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failure vs a bare-except bug in Fuyu tensor conversion." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer/model families and different failure modes." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLuke tokenizer AttributeError is unrelated to Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Base-model-prefix loading regression vs rope-param validation TypeError; distinct code paths." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBird tokenizer special-token registration bug vs Qwen recurrent-state reset bug." + }, + { + "left": "issue:29870", + "right": "issue:34689", + "accept": false, + "reason": "Bart generate decoder mask bug vs Llama vision model loading break; no shared underlying fix." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 processor import failure vs flash_attention sink-less crash; different layers." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch vs DeepSeekV3MoE implementation divergence; not the same issue." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "FineGrainedFP8 model-load failure vs quantized model saving failure." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported feature-extractor arg vs video-input regression; different subsystems." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures vs image processor void-segmentation handling." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBird mask-token issue is unrelated to FastSpeech2ConformerTokenizer loading." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Exception handling for protobuf decode errors vs leftover dummy class cleanup; different fixes." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export exportability issue vs Wav2Vec2 save/tokenization failure; not the same code path." + }, + { + "left": "issue:45491", + "right": "issue:46693", + "accept": false, + "reason": "Both involve FA2/long-context generation, but one is NaN correctness and the other is a performance regression; not the same bug." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Processor import error vs transformers serve missing _tokenizer attribute; unrelated failures." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Dummy Conv2d cleanup vs SigLIP2 nonsensical outputs; no shared underlying defect." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export guard/support issue vs flash_attention attribute crash; unrelated." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and CLIP bus error are unrelated." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count runtime limit vs paged generate warning; different concerns." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy class cleanup vs RT-DETR memory release bug; distinct issues." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM vs resume_from_checkpoint key mismatch; not the same change." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepSeek implementation divergence vs D-FINE auxiliary-loss behavior; unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Path-trailing-slash import warning vs gradient checkpointing eval-mode restriction." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma4 import failure and protobuf error masking are separate tokenizer/processor problems." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error messaging vs embedding reinitialization after resize; different functionality." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 34, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:42915", + "issue:43066", + "issue:43334", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43646", + "issue:43761", + "issue:43824", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44077", + "issue:44186", + "issue:44279", + "issue:44387", + "issue:44442", + "issue:44462", + "issue:44464", + "issue:44514", + "issue:44743", + "issue:44811", + "issue:44898", + "issue:44977", + "issue:45072", + "issue:45245", + "issue:45250", + "issue:45381", + "issue:45412", + "issue:45491", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45834", + "issue:45859", + "issue:46121", + "issue:46369", + "issue:46399", + "issue:46640", + "issue:46693" + ], + "soft_pairs": [ + "issue:43066|issue:44462", + "issue:43421|issue:45491", + "issue:43994|issue:45412", + "issue:43575|issue:44387", + "issue:45412|issue:46369", + "issue:44442|issue:45072", + "issue:43526|issue:45834", + "issue:44464|issue:45584", + "issue:42915|issue:43931", + "issue:43526|issue:46693", + "issue:43824|issue:44279", + "issue:43994|issue:46369", + "issue:27957|issue:29127", + "issue:27957|issue:33290", + "issue:43526|issue:44186", + "issue:33290|issue:34689", + "issue:33290|issue:35141", + "issue:46121|issue:46399", + "issue:43334|issue:43531", + "issue:45250|issue:46640", + "issue:45588|issue:45600", + "issue:43381|issue:43873", + "issue:43646|issue:44077", + "issue:43761|issue:45859", + "issue:44514|issue:45381", + "issue:44977|issue:45245", + "issue:44743|issue:44811", + "issue:44442|issue:44898" + ], + "prepared_review_unit_hash": "bf58f0ed4c7f5bf4f5528a4ff2afc89bc6a76ca182a9675e5b768002dc6f956a", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35299, + "estimated_input_tokens": 8825, + "estimated_eval_tokens": 17906 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose collection of unrelated Transformers issues with a few flash-attention/Qwen-adjacent items; none of the soft pairs look like true duplicates with the evidence shown.", + "confidence": 0.18, + "canonical_issue_reason": "issue:46640 is the best umbrella candidate: it is the most general flash-attention-2 report in the set and sits closest to the small FA2-related subgroup, though the overall cluster is too heterogeneous to treat as one duplicate family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46640 is the most representative issue for this set because it is broad, current, and closest to the only visible thematic thread (Flash Attention 2).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Both concern tokenizer loading behavior, but one is about decoder type in v5 and the other about ignoring tokenizer.json; different bugs and code paths." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "One is about runtime special-token/post-processor updates, the other about Gemma3 NaNs from sliding-window all-padding cases; unrelated failures." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "AutoModel/pipeline wrong results for SigLIP2 is unrelated to RT-DETR memory not being released." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both mention OOM, but one is tensor-parallel model loading and the other is int4 quantization reserved-memory growth; not the same bug." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "Both are memory-related, but one is a GC/leak issue in RT-DETR and the other is model-call memory overflow for sam3-litetext; too different." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "FastSpeech2 tokenizer loading and bfloat16 dtype mismatches are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels and Kosmos2.5 OCR index errors are different image-processing bugs." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is unrelated to Whisper empty-transcription handling after align_special_tokens." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and Qwen3-VL weight-shape mismatch are separate model/loading problems." + }, + { + "left": "issue:43526", + "right": "issue:46693", + "accept": false, + "reason": "reduce_labels behavior in an image processor is unrelated to an FA2 long-context performance regression." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "An import error for Qwen2_5_VL and a generic dependency issue are not the same concrete bug." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "SigLIP2 bad outputs and sam3-litetext memory overflow do not share a code path." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta and LayoutLMv3 box-validation error messaging are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "FA2 support for XLMRoberta and DeepSpeed AdaFactor OOM are different areas and causes." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and LayoutLMv2Tokenizer padding/NER crashes are separate tokenizer/processor bugs." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "AdaFactor/deepspeed OOM is unrelated to Llama 3.2 Vision model loading breakage." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "OOM in training with AdaFactor is not the same as embeddings being reinitialized during post_init." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "Both touch rope/rotary machinery, but one is a type-conversion validation issue and the other is a missing decorator registration crash." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id on Qwen3-VL text config and Qwen3-MoE sliding_window behavior are distinct issues." + }, + { + "left": "issue:45250", + "right": "issue:46640", + "accept": false, + "reason": "Both are about Flash Attention 2 in name only; one is a generic placeholder title and the other is a specific performance regression." + }, + { + "left": "issue:45588", + "right": "issue:45600", + "accept": false, + "reason": "Flash-attention AttributeError on sink-less models and removed Sam3LiteText auto-mapping refs are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode and quantization offloading behavior are different features and failures." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "A broad Transformers 5.0 custom initialization regression is not the same as patchtsmixer's optional post_init constraint." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen3_5Moe rotary embedding CPU-offload incompatibility are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Batched apply_chat_template padding=False crash and Qwen2.5-VL video vision_position_ids bug are different Qwen2.5-VL failure modes." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3.5 flash-attention generation issues and category-cardinality runtime limits are unrelated." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Qwen3.5 recurrent-state caching and Whisper batch_decode skip_special_tokens behavior are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "FastSpeech2 tokenizer loading and Perceiver interpolation failures are different model/tokenizer bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 35, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:27957", + "issue:29870", + "issue:42175", + "issue:42490", + "issue:42907", + "issue:43232", + "issue:43262", + "issue:43316", + "issue:43388", + "issue:43493", + "issue:43525", + "issue:43643", + "issue:43701", + "issue:43720", + "issue:43749", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44188", + "issue:44190", + "issue:44206", + "issue:44442", + "issue:44479", + "issue:44496", + "issue:44743", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:44987", + "issue:45003", + "issue:45137", + "issue:45216", + "issue:45325", + "issue:45440", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:45907", + "issue:46178", + "issue:46234", + "issue:46506", + "issue:46697", + "issue:46829", + "issue:46932" + ], + "soft_pairs": [ + "issue:43262|issue:44206", + "issue:42490|issue:44743", + "issue:45003|issue:45440", + "issue:43720|issue:44442", + "issue:42175|issue:46234", + "issue:43493|issue:43827", + "issue:44077|issue:44479", + "issue:44062|issue:45907", + "issue:27957|issue:29870", + "issue:43316|issue:43525", + "issue:45325|issue:46697", + "issue:43873|issue:44871", + "issue:43749|issue:43931", + "issue:44855|issue:45137", + "issue:43643|issue:44913", + "issue:43316|issue:45759", + "issue:44190|issue:44496", + "issue:44442|issue:45803", + "issue:43761|issue:44079", + "issue:43761|issue:44188", + "issue:44987|issue:46178", + "issue:43931|issue:45216", + "issue:43232|issue:43388", + "issue:46829|issue:46932", + "issue:43701|issue:46506", + "issue:42907|issue:45698" + ], + "prepared_review_unit_hash": "e31b167131d4c1aeb53f09233e47e29cf50f6f3472aeb94e9a283c73ee66d527", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35810, + "estimated_input_tokens": 8953, + "estimated_eval_tokens": 18162 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous; the soft pairs share at most broad subsystem vocabulary, not the same underlying bug. No PRs are present, and there is no clear canonical issue for the whole cluster.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sampling-rate mismatch vs a LasrFeatureExtractor center-arg crash; unrelated code paths and failures." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Path-with-trailing-slash import warning is unrelated to recurrent-state reset in modular_qwen3_5 cache handling." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils is a different problem from DeepSeekV3MoE behavioral divergence." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading failure and FastSpeech2ConformerTokenizer loading failure affect different loaders and models." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Missing TensorFlow in a torch install extra is not the same as a missing import_utils availability helper." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation drift is unrelated to documentation still referencing removed pipelines." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "patchtsmixer post_init allowance is a model-config issue, not the Qwen video-input regression." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "AddedToken special-argument conflict and list_repo_files revision forwarding are unrelated API bugs." + }, + { + "left": "issue:27957", + "right": "issue:29870", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to Bart generate() using decoder_attention_mask before assignment." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and missing pad_token_id on Llama4Config are different config-field problems." + }, + { + "left": "issue:45325", + "right": "issue:46697", + "accept": false, + "reason": "Qwen2.5-VL rope-index scaling bug and Swin2SR over-padding bug are different image/video preprocessing issues." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior does not match the Gemma-3 eos_token_id configuration inconsistency." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and Qwen3-VL weight-shape mismatch are distinct loading failures." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 decorator parsing error is unrelated to DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "trust_remote_code field loss on AutoConfig reload is unrelated to GPTNeoX rotary_pct not persisting." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Both mention config handling, but one is Gemma3TextConfig API shape and the other is AutoModelForCausalLM.from_config unwrapping text_config; not the same bug." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in an example script is unrelated to an unrecognized model_type in config.json." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure and FuyuBatchFeature's bare except are different parts of the stack and different symptoms." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression is unrelated to ModelOutput key assignment when a key was previously None." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Hidden_states regression in CLIPVisionModel and attention-kernel divergence under torch.compile are different bugs." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "A model-loading failure for physical-intelligence/fast is unrelated to adding tensorflow_text to BACKENDS_MAPPING." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Qwen3-VL shape mismatch on load and Qwen3.5 save_pretrained checkpoint corruption are different lifecycle stages and causes." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs sync bug and gather_for_metrics truncating tuple labels are unrelated training/runtime issues." + }, + { + "left": "issue:46829", + "right": "issue:46932", + "accept": false, + "reason": "VideoPrism hidden_states/output bug is unrelated to Gemma4 quantized vision/audio embedder dtype casting." + }, + { + "left": "issue:43701", + "right": "issue:46506", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is unrelated to CPU/CUDA divergence in PPDocLayoutV3." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals and loading the wrong custom module after save_pretrained are different persistence issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 36, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:27957", + "issue:28282", + "issue:32090", + "issue:42617", + "issue:43295", + "issue:43316", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43650", + "issue:43701", + "issue:43716", + "issue:43742", + "issue:43761", + "issue:43825", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44464", + "issue:44479", + "issue:44496", + "issue:44534", + "issue:44610", + "issue:44792", + "issue:44871", + "issue:44913", + "issue:45584", + "issue:45632", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46202", + "issue:46234", + "issue:46291", + "issue:46365", + "issue:46369", + "issue:46421", + "issue:46424", + "issue:46821", + "issue:46894" + ], + "soft_pairs": [ + "issue:42617|issue:43575", + "issue:43575|issue:43742", + "issue:43550|issue:43701", + "issue:27957|issue:32090", + "issue:43606|issue:43701", + "issue:25251|issue:28282", + "issue:46234|issue:46291", + "issue:43295|issue:43825", + "issue:16998|issue:22355", + "issue:43295|issue:44871", + "issue:44871|issue:45584", + "issue:44464|issue:46421", + "issue:44077|issue:44534", + "issue:43316|issue:43716", + "issue:43316|issue:46821", + "issue:44871|issue:46050", + "issue:43646|issue:44913", + "issue:43452|issue:44496", + "issue:45803|issue:46365", + "issue:44162|issue:44792", + "issue:43404|issue:43441", + "issue:45632|issue:46424", + "issue:44242|issue:46894", + "issue:45803|issue:46170", + "issue:44610|issue:46369", + "issue:43650|issue:46202", + "issue:43761|issue:44479" + ], + "prepared_review_unit_hash": "02f2943582db37fb55673de99e02789487e6d5dcbcb6ce1e4a9956bbfcab558a", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 34962, + "estimated_input_tokens": 8741, + "estimated_eval_tokens": 17738 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose set of unrelated Transformers issues with only superficial title overlap. None of the soft edges look like true duplicates or the same fixable code-path change, so no pair should be merged.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 44871 is the most concrete, self-contained regression report and the most central node among the soft links, but the surrounding items are heterogeneous rather than duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "44871 is the best representative issue because it describes a specific, actionable bug with a narrow symptom; it is more triage-friendly than the more generic or broader reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "Different failures: running a 3d_parallel example vs TP-loading Qwen2 causing OOM." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "Different model-loading problems with different symptoms and likely different causes." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "Torch.compile/SDPA failure is unrelated to a resume_from_checkpoint key mismatch." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to a Trainer broadcast NoneType error." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch for bark-small is unrelated to checkpoint key naming." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k output shape regression is unrelated to an ImportError about missing PyTorch." + }, + { + "left": "issue:46234", + "right": "issue:46291", + "accept": false, + "reason": "Missing utility symbol and kernels import-version failure are separate import-time issues." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Processor/tokenizer regression is unrelated to an incorrect pipeline error message about translation tasks." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "A model_max_length question and a missing transformers.onnx import are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "One is a processor/tokenizer API regression; the other is an eos_token_id config inconsistency." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Gemma eos_token_id mismatch is unrelated to Whisper empty-transcription generation failure." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward is not the same as lazy cache initialization recompilation." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "PatchTSMixer post_init policy and non-persistent buffer corruption are different bugs." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy is unrelated to image preprocessor dtype mismatch." + }, + { + "left": "issue:43316", + "right": "issue:46821", + "accept": false, + "reason": "Config API mismatch and TimesFM window_size AttributeError are different model-specific issues." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Gemma eos_token_id inconsistency is unrelated to quantized sinq save/load breakage." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Custom model init breakage in Transformers 5 is not the same as rotary_pct not persisting on reload." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "gguf_file handling and unrecognized model config are different loading-path failures." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Bare except in Fuyu tensor conversion is unrelated to the Trainer import issue involving datasets." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breakage and a janus image-generation test failure are not the same bug." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "lm_head weight tying in Mistral3 and FlashAttention failures in Ministral-3 are separate model regressions." + }, + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Trust-remote-code cache path collisions are unrelated to max_cache_length being honored only for quantized cache." + }, + { + "left": "issue:44242", + "right": "issue:46894", + "accept": false, + "reason": "MoE load-balancing loss omission is not the same as a CohereASR double-shift training-loss bug." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Fuyu's bare except problem is unrelated to Windows/ROCm torch.distributed import dependency issues." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "OmDet-Turbo input-size mismatch and sam3-litetext memory overflow are different model problems." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "A vague data request is not a duplicate of a Gemma roundtrip failure." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVision hidden_states regression and video-input regression affect different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 37, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:35141", + "issue:41950", + "issue:42490", + "issue:43065", + "issue:43352", + "issue:43388", + "issue:43475", + "issue:43504", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43572", + "issue:43673", + "issue:43688", + "issue:43827", + "issue:43994", + "issue:44190", + "issue:44514", + "issue:44534", + "issue:44617", + "issue:44734", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44869", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45127", + "issue:45200", + "issue:45245", + "issue:45406", + "issue:45542", + "issue:45588", + "issue:45600", + "issue:45632", + "issue:45698", + "issue:46082", + "issue:46178", + "issue:46234", + "issue:46421", + "issue:46617", + "issue:46640", + "issue:46697", + "issue:46832", + "issue:46858", + "issue:46932" + ], + "soft_pairs": [ + "issue:43540|issue:44514", + "issue:43504|issue:45698", + "issue:45542|issue:46234", + "issue:43388|issue:43688", + "issue:43065|issue:44617", + "issue:44792|issue:44977", + "issue:44792|issue:45600", + "issue:43572|issue:46178", + "issue:45127|issue:45200", + "issue:43673|issue:44743", + "issue:43352|issue:46640", + "issue:41950|issue:43475", + "issue:44190|issue:44964", + "issue:44534|issue:44913", + "issue:43525|issue:43531", + "issue:44734|issue:45406", + "issue:35141|issue:42490", + "issue:43827|issue:43994", + "issue:45245|issue:46858", + "issue:44869|issue:46832", + "issue:45072|issue:46932", + "issue:46082|issue:46697", + "issue:45632|issue:46617", + "issue:44855|issue:45588", + "issue:43673|issue:46421" + ], + "prepared_review_unit_hash": "c73b1a833acc84cb07da6c6d4e64b98719f775ca1e5d50a3526bb7f5f7b9723a", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 25, + "serialized_chars": 35381, + "estimated_input_tokens": 8846, + "estimated_eval_tokens": 17948 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft pairs mostly share only broad subsystem vocabulary (cache, loading, flash-attention, processors) but not the same concrete bug or change. I would reject every soft edge and avoid collapsing these items into one duplicate cluster.", + "confidence": 0.98, + "canonical_issue_reason": "No strong canonical issue: the items span unrelated model, cache, processor, config, server, and docs problems, so there is no single duplicate target that represents the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue; the set is not a true duplicate cluster, so any single issue would be an arbitrary representative rather than a meaningful canonical.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Different bugs: one is a video-input processing failure in Qwen3OmniMoe, the other is a batched chat-template crash in Qwen2_5_VLProcessor with padding=False." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Both concern loading, but one is a BEiT preset legacy-field failure and the other is a wrong custom-module reload after save_pretrained; different root causes." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "Unrelated optional-dependency issues: tf backend ambiguity vs a missing import_utils availability helper for falcon mamba." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Evaluation label collation in gather_for_metrics is unrelated to auxiliary-loss normalization in OLMoE/GPT Oss." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Same model family only superficially; a dummy Conv2d in Sam3PixelDecoder is not the same bug as Sam3Video CUDA OOM." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation failure are different models and failure modes." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "A Janus generation test failure is unrelated to stale auto_mappings references to removed Sam3LiteText configs." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLm pad_token_id regression and a tensorflow_text backend mapping error are different config/import problems." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings is not the same as mm_token_type_ids defaulting for text-only fine-tuning." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache, but one is missing cache in chunked_prefill and the other is recurrent states resetting with seq_len>1; distinct concrete failures." + }, + { + "left": "issue:43352", + "right": "issue:46640", + "accept": false, + "reason": "One says NemotronHForCausalLM does not support Flash Attention 2.0 yet; the other is a separate generic flash-attention issue." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification processor selection and a missing fpn_position_embeddings attribute in Sam3 vision output are unrelated." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading in a training script is not the same bug as loading Phi-4-multimodal-instruct with the latest transformers." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk initialization and GPTNeoX rotary_pct reload regression are unrelated configuration/runtime issues." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is unrelated to the Qwen3-MoE sliding_window issue." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "A serve KV-cache indexing crash and a Gemma4Processor missing _tokenizer crash are different server-side bugs." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Embedding reinitialization after resize and trailing-slash local-model import warnings do not describe the same problem." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing pipeline() after removals is not the same as SigLIP2 producing nonsense with AutoModel and pipeline." + }, + { + "left": "issue:45245", + "right": "issue:46858", + "accept": false, + "reason": "A category-count overflow and a device-tensor cache length arithmetic bug are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:46832", + "accept": false, + "reason": "Whisper timestamp decoding with trailing replacement characters is unrelated to RT-DETR feature-level mismatch in forward()." + }, + { + "left": "issue:45072", + "right": "issue:46932", + "accept": false, + "reason": "These are different dtype issues in different model paths: bfloat16 inference mismatches vs quantized vision/audio embedder casting." + }, + { + "left": "issue:46082", + "right": "issue:46697", + "accept": false, + "reason": "LlamaConfig head_dim validation and Swin2SR over-padding are unrelated bugs." + }, + { + "left": "issue:45632", + "right": "issue:46617", + "accept": false, + "reason": "Both involve trust_remote_code caching, but one is a leaf-directory cache collision and the other is a FileNotFoundError from cache loading; not the same root cause." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "An import/torch.jit parsing error in DebertaV2Model is unrelated to the flash_attention.py s_aux=None AttributeError." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Missing GenerationMixin cache during chunked_prefill and lazy cache initialization causing recompilation are distinct cache-related problems, not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 38, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:29870", + "issue:35141", + "issue:39692", + "issue:41720", + "issue:43065", + "issue:43295", + "issue:43335", + "issue:43526", + "issue:43550", + "issue:43644", + "issue:43931", + "issue:44060", + "issue:44077", + "issue:44112", + "issue:44190", + "issue:44242", + "issue:44462", + "issue:44464", + "issue:44479", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44737", + "issue:44779", + "issue:44855", + "issue:44877", + "issue:44913", + "issue:44933", + "issue:45127", + "issue:45245", + "issue:45276", + "issue:45290", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45563", + "issue:45584", + "issue:45600", + "issue:45803", + "issue:45907", + "issue:46369", + "issue:46489", + "issue:46655", + "issue:46697", + "issue:46752", + "issue:46897" + ], + "soft_pairs": [ + "issue:43931|issue:44060", + "issue:29870|issue:35141", + "issue:44617|issue:45584", + "issue:44242|issue:46897", + "issue:43644|issue:44913", + "issue:43335|issue:44855", + "issue:45563|issue:45584", + "issue:25251|issue:27957", + "issue:43550|issue:44190", + "issue:44933|issue:45362", + "issue:44077|issue:44877", + "issue:41720|issue:45245", + "issue:45381|issue:46369", + "issue:44479|issue:44734", + "issue:43335|issue:44190", + "issue:43065|issue:44610", + "issue:39692|issue:44190", + "issue:45127|issue:45357", + "issue:44464|issue:45563", + "issue:45381|issue:46655", + "issue:44779|issue:46489", + "issue:44617|issue:46369", + "issue:43295|issue:45276", + "issue:44112|issue:45600", + "issue:43526|issue:46697", + "issue:45290|issue:46752", + "issue:44462|issue:45907", + "issue:44737|issue:45803", + "issue:45276|issue:45584" + ], + "prepared_review_unit_hash": "3049808d6e407ea7a299d7545ab3e8e369a97ed161ca6808deec875746beffe9", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36526, + "estimated_input_tokens": 9132, + "estimated_eval_tokens": 18520 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated transformer/model bug reports that only share broad subsystem keywords. I found one plausible duplicate pair in the DeepSeek tokenizer regression on v5+; the rest are distinct issues and should stay separate.", + "confidence": 0.89, + "canonical_issue_reason": "issue:46489 is the clearest and most specific representative of the only true near-duplicate pair: it describes the DeepSeek tokenizer v5+ regression as a gap in the earlier fix, making it the best canonical issue for that bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46489 is the best overall issue here because it pinpoints a concrete, actionable tokenizer regression with an explicit tie to the prior fix path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Different models and different failures: Qwen3-VL weight-shape mismatch vs Qwen3-Next tied-weights warning." + }, + { + "left": "issue:29870", + "right": "issue:35141", + "accept": false, + "reason": "Separate bugs: BART generate() hits an UnboundLocalError, while the other is output-embedding reinitialization after resize_token_embeddings." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated symptoms and code paths: Sam3Video CUDA OOM vs Whisper empty-transcription generation crash." + }, + { + "left": "issue:44242", + "right": "issue:46897", + "accept": false, + "reason": "Both concern training losses, but they are different model-specific bugs with different root causes." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Different serialization/config problems: non-persistent buffers junk vs GPTNeoXConfig rotary_pct not persisting on reload." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "One is a SwitchTransformers config construction bug; the other is a Python 3.13 import/parsing failure in DebertaV2Model." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "Different generate() issues: a stale warning about num_return_sequences vs a crash on empty Whisper transcription." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "No shared bug: pipeline top_k nesting regression vs XLMRoberta Flash Attention 2 support issue." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Completely different failures: torch.compile/SDPA on Bamba vs local dataset loading in an example script." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Unrelated: missing import from image_utils vs Qwen3.5-35B chat crash." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-related, but one is about patchtsmixer post_init allowance and the other about strict config loading for granite_speech." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Different classes of bugs: Qwen3 auto device mapping cuda assert vs category-count limit RuntimeError." + }, + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Different issues: Qwen video-position IDs regression vs Sam3-LiteText memory overflow." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Both hit generation/video-serving paths, but the concrete bugs differ: video-input regression in Qwen models vs KV-cache continuation indexing in serve." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "Unrelated bugs: sparse-layer creation in SwitchTransformers vs local dataset loading failure." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Different model/component bugs: a dummy Conv2d in Sam3PixelDecoder vs OmDet-Turbo processor/model size mismatch." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "Docs/example errors in SigLIP2 are unrelated to the local dataset loading problem." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "Different mechanisms: LoRA merge with extended vocab collapse vs save_pretrained writing wrong visual encoder keys." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Different generation bugs: compiled-forward chunked generation inconsistency vs a stale warning for num_return_sequences." + }, + { + "left": "issue:45381", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-related, but the underlying bugs differ: wrong Qwen video position IDs vs over-padding in SmolVLM video preprocessing." + }, + { + "left": "issue:44779", + "right": "issue:46489", + "accept": true, + "reason": "Same underlying bug: DeepSeek tokenizer output is wrong on transformers v5+, and 46489 is a follow-up describing the gap in the earlier fix." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Both mention Sam3 memory pressure, but they are different models and different reports (OOM vs memory overflow)." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Distinct resize/tokenizer regressions with different affected code paths and symptoms." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both are CI failures, but they stem from unrelated causes: stale GraniteSpeech device override test vs removed Sam3LiteText config references." + }, + { + "left": "issue:43526", + "right": "issue:46697", + "accept": false, + "reason": "Different image-processing bugs: BEiT label reduction vs Swin2SR over-padding." + }, + { + "left": "issue:45290", + "right": "issue:46752", + "accept": false, + "reason": "Both are chat-template crashes, but the triggers differ: assistant tool-call message with no content vs passing an empty message list." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Separate repository/loader bugs: AutoTokenizer ignoring tokenizer.json vs list_repo_files not forwarding revision." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "Unrelated: XLNet device placement bug vs FuyuBatchFeature swallowing exceptions." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Different domains entirely: Gemma4 embedding resize bug vs Whisper empty-transcription failure." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet. The DeepSeek tokenizer pair (issue:44779 / issue:46489) is the only clearly plausible duplicate-like match, and the other soft-edge verdicts are appropriately conservative given the titles and available evidence. I didn\u2019t see an overstatement of evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 39, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:29870", + "issue:32090", + "issue:35141", + "issue:36246", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42796", + "issue:43295", + "issue:43404", + "issue:43531", + "issue:43540", + "issue:43582", + "issue:43643", + "issue:43650", + "issue:43673", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44295", + "issue:44496", + "issue:44521", + "issue:44734", + "issue:44938", + "issue:44964", + "issue:44987", + "issue:45003", + "issue:45083", + "issue:45216", + "issue:45335", + "issue:45381", + "issue:45542", + "issue:45600", + "issue:45632", + "issue:46082", + "issue:46121", + "issue:46234", + "issue:46424", + "issue:46506", + "issue:46858", + "issue:46932" + ], + "soft_pairs": [ + "issue:43540|issue:44734", + "issue:44938|issue:45542", + "issue:44521|issue:45381", + "issue:25251|issue:29870", + "issue:43673|issue:46858", + "issue:43854|issue:44964", + "issue:45632|issue:46234", + "issue:43643|issue:44496", + "issue:44246|issue:45003", + "issue:44964|issue:45600", + "issue:44112|issue:44964", + "issue:42175|issue:44279", + "issue:27957|issue:35141", + "issue:46082|issue:46121", + "issue:42796|issue:44521", + "issue:36246|issue:41628", + "issue:43404|issue:45216", + "issue:32090|issue:35141", + "issue:36246|issue:38617", + "issue:43531|issue:45600", + "issue:41950|issue:44016", + "issue:44295|issue:46506", + "issue:43650|issue:44220", + "issue:43673|issue:46424", + "issue:43976|issue:44987", + "issue:43582|issue:46932", + "issue:43295|issue:45335", + "issue:45083|issue:46506", + "issue:44016|issue:44220" + ], + "prepared_review_unit_hash": "6f538696449425b441e94c5620666effcb163fa47ec7ae9ed8df606f3d99a96b", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36493, + "estimated_input_tokens": 9124, + "estimated_eval_tokens": 18504 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the soft-linked issues share broad themes like model loading, multimodal inputs, caching, or import errors, but they do not appear to describe the same underlying bug/change.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue stands out because the cluster spans many unrelated defects across different subsystems and models.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong representative for the whole cluster; the items are too mixed to anchor on one canonical bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Different code paths: Qwen3OmniMoe video-input processing vs transformers serve KV-cache continuation indexing." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Unrelated failures: Python 3.14 import/load problem vs missing TensorFlow backend when only tensorboard is installed." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal/Qwen-related, but one is assistant mask generation and the other is video position IDs; not the same bug." + }, + { + "left": "issue:25251", + "right": "issue:29870", + "accept": false, + "reason": "Pipeline output nesting change and BART generate decoder_attention_mask error are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:46858", + "accept": false, + "reason": "Both mention cache, but one is missing cache in chunked_prefill while the other is a device-tensor seq-length arithmetic issue." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Both are model-load regressions, but they affect different models and likely different root causes." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Trust-remote-code cache-path collision and a missing import_utils symbol are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Both concern config/model loading, but one is missing returned fields with trust_remote_code and the other is an unrecognized model lacking model_type." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import-time performance complaint vs unsafe sys.modules access; not the same bug." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Different problems: Phi-4 multimodal model loading vs stale auto_mappings references breaking CI." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI test flakiness in GraniteSpeech is unrelated to Phi-4 multimodal loading failure." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Dependency/backend packaging issue and a general dependency issue are too vague and not the same defect." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to embedding reinitialization after resize_token_embeddings." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "LlamaConfig head_dim validation and rope-param conversion TypeError are different configuration bugs." + }, + { + "left": "issue:42796", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer BPE-marker handling and multimodal assistant masks are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they refer to different missing names and modules." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Different regression types and models: untied lm_head in Mistral3ForConditionalGeneration vs incorrect saved checkpoint for Qwen3.5." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer broadcast TypeError and token-embedding resize reinitialization do not share a code path." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Missing Qwen2_5_VLImageProcessor and missing layer_type_validation are unrelated import/export issues." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and stale config references in auto_mappings are unrelated." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline bug and notebook syntax error are not duplicates." + }, + { + "left": "issue:44295", + "right": "issue:46506", + "accept": false, + "reason": "Position_ids buffer access error and CPU/CUDA result divergence are different defects." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "Placeholder/data request issue and _torch_extract_fbank_features bug are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Both mention cache configuration, but they concern different semantics and code paths." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python-version compatibility failure and a model-loading failure for physical-intelligence/fast are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:46932", + "accept": false, + "reason": "AppleSilicon TypeError in allocator warmup and Gemma4 dtype-casting fix are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Processor/tokenizer regression and t5gemma decoder embedding resize bug are different issues." + }, + { + "left": "issue:45083", + "right": "issue:46506", + "accept": false, + "reason": "Feature-extractor length helper behavior in qwen3_omni_moe and PPDocLayoutV3 CPU/CUDA mismatch are unrelated." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and fbank feature extraction bug are not the same problem." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 40, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:39692", + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43479", + "issue:43526", + "issue:43577", + "issue:43673", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44279", + "issue:44327", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44521", + "issue:44683", + "issue:44779", + "issue:44792", + "issue:44933", + "issue:45042", + "issue:45216", + "issue:45278", + "issue:45335", + "issue:45375", + "issue:45405", + "issue:45446", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45812", + "issue:46082", + "issue:46123", + "issue:46369", + "issue:46399", + "issue:46424", + "issue:46512", + "issue:46515", + "issue:46693", + "issue:46752" + ], + "soft_pairs": [ + "issue:44496|issue:45042", + "issue:43526|issue:43901", + "issue:43381|issue:45593", + "issue:42907|issue:43441", + "issue:44683|issue:45446", + "issue:43901|issue:44496", + "issue:43673|issue:44336", + "issue:43479|issue:45375", + "issue:43976|issue:44279", + "issue:44496|issue:45759", + "issue:44016|issue:44062", + "issue:44792|issue:46424", + "issue:43441|issue:45588", + "issue:43824|issue:44933", + "issue:44279|issue:46515", + "issue:44521|issue:46693", + "issue:45335|issue:45584", + "issue:44327|issue:46752", + "issue:46123|issue:46369", + "issue:43299|issue:45216", + "issue:46082|issue:46399", + "issue:44279|issue:45405", + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:44075|issue:44368", + "issue:43577|issue:45216", + "issue:46082|issue:46512", + "issue:43976|issue:45405", + "issue:44779|issue:45812", + "issue:44351|issue:45278" + ], + "prepared_review_unit_hash": "0bd075336304f2516e0af688e7e2281a507787da8057cf7332192ed9ae39dae7", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36769, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: most items are isolated model/config/regression reports, so there is no strong duplicate core. I treated every soft edge as a false match because the paired artifacts do not describe the same concrete bug or change.", + "confidence": 0.63, + "canonical_issue_reason": "Issue 45278 is the broadest regression report in the set, describing multiple import failures after an upgrade. It is the closest thing to an umbrella issue, even though the cluster is mostly unrelated one-off bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the best stand-alone representative because it captures a wider API/import breakage rather than a single model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Different failures in different areas: model loading/config resolution vs PIL image processor torchvision dependency." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "One is a BeitImageProcessorFast label-processing bug; the other is outdated docs for text classification pipeline behavior." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing eval-mode behavior and D-FINE denoising losses are unrelated code paths." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Both mention Ministral/devstral, but one is about saving dequantized models while the other is a FlashAttention load/runtime failure." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both touch flex_attention, but one is a torch>=2.9 compiled-kernel failure and the other is a version-check/import bug for AuxRequest." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Documentation mismatch for return_all_scores is unrelated to an unrecognized-model config error." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Generation cache handling during chunked prefill is unrelated to ANSI coloring in loading_report." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Both are config bugs, but they affect different multimodal model classes and different fields/defaulting behavior." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python version support breakage is not the same underlying problem as a generic dependency issue." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Different multimodal model-construction failures: one rejects a model config, the other fails to unwrap text_config in from_config." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and AddedToken keyword conflict are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "A janus image-generation test failure is not the same as cache-length handling in quantized cache." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both involve flash attention, but the concrete failures differ: Ministral-3 runtime breakage vs s_aux=None AttributeError for sink-less models." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Missing Qwen2_5_VL export and a nonexistent image_utils import are separate import problems." + }, + { + "left": "issue:44279", + "right": "issue:46515", + "accept": false, + "reason": "A generic transformers dependency issue is not the same as an explicit torch 2.4.1 incompatibility." + }, + { + "left": "issue:44521", + "right": "issue:46693", + "accept": false, + "reason": "Assistant-mask generation in multimodal chat templates is unrelated to FA2 long-context performance regression." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize propagation in t5gemma is unrelated to Whisper empty-transcription generation after align_special_tokens." + }, + { + "left": "issue:44327", + "right": "issue:46752", + "accept": false, + "reason": "QA span decoding top-k bounds and empty apply_chat_template input are different runtime errors." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "A partial-batch flag bug in MaskGenerationPipeline is unrelated to memory overflow in sam3-litetext." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Qwen3VL Moe loading regression and Qwen3.5 save_pretrained checkpoint corruption are different code-path failures." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "LlamaConfig head_dim validation and Qwen3.5 kernel registration error are unrelated." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "A vague dependency issue is not the same as an explicit MIN_PEFT_VERSION pin to an unreleased version." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Custom model processor.tokenizer breakage and Deepseek tokenizer output regression are both tokenizer-related but not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "SigLIP2 docs example errors are unrelated to a text-classification pipeline documentation mismatch." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Unused SGD optimizer args and a tie_word_embeddings warning are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation and Qwen3.5 save_pretrained checkpoint correctness are different regressions." + }, + { + "left": "issue:46082", + "right": "issue:46512", + "accept": false, + "reason": "Config validation for Llama head_dim has nothing to do with stale model.config.dtype after from_config(dtype=...)." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python support breakage and PEFT version pinning are separate packaging issues." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions affecting different model families and code paths." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A specific HybridCache import failure is not clearly the same underlying bug as the broader multi-import regression report." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 41, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:33290", + "issue:41950", + "issue:42175", + "issue:42796", + "issue:42994", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43650", + "issue:43653", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44242", + "issue:44261", + "issue:44263", + "issue:44291", + "issue:44336", + "issue:44361", + "issue:44462", + "issue:44466", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:45020", + "issue:45081", + "issue:45103", + "issue:45200", + "issue:45276", + "issue:45357", + "issue:45372", + "issue:45397", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712", + "issue:45823", + "issue:45834", + "issue:45859", + "issue:45907", + "issue:46209", + "issue:46851" + ], + "soft_pairs": [ + "issue:42796|issue:43650", + "issue:41950|issue:45020", + "issue:43299|issue:45357", + "issue:44466|issue:45276", + "issue:43299|issue:43854", + "issue:43653|issue:45593", + "issue:44242|issue:45200", + "issue:41950|issue:45823", + "issue:43650|issue:46209", + "issue:44462|issue:45081", + "issue:43479|issue:43525", + "issue:44016|issue:44291", + "issue:44683|issue:45542", + "issue:44361|issue:45593", + "issue:45081|issue:45907", + "issue:43441|issue:45834", + "issue:33290|issue:45656", + "issue:45103|issue:45372", + "issue:44336|issue:46851", + "issue:42994|issue:43317", + "issue:45859|issue:46851", + "issue:44623|issue:44749", + "issue:44261|issue:44263", + "issue:45600|issue:45712", + "issue:42175|issue:43976", + "issue:44671|issue:44779", + "issue:44623|issue:45397" + ], + "prepared_review_unit_hash": "49e96a24d55140ae976e95ec2309e3544ea8dd520c052768b4609a725ce34525", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35217, + "estimated_input_tokens": 8805, + "estimated_eval_tokens": 17866 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The proposed soft pairs are overwhelmingly false positives: they mostly span unrelated bug reports across tokenization, loading, quantization, DeepSpeed, model configs, and CI. There is no real duplicate cluster here.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue stands out; the items are heterogeneous and do not share one underlying bug or code path.", + "canonical_pr_reason": null, + "best_issue_reason": "If one issue had to anchor the set, #44623 is the broadest standalone API regression, but it still does not represent the rest of the items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42796", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer BPE-marker handling vs an unrelated data/title-only issue; no shared code path." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video pipeline processor lookup vs remote_code compatibility regressions; different subsystems." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "MoE model loading failure vs save_pretrained visual-key serialization; different operations." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Weight serialization inconsistency vs resize_token_embeddings propagation; not the same bug." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different models and failure modes: Qwen3VL loading vs GLM-4.7 unit-test loading." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer special-token registration vs D-FINE auxiliary-loss logic; unrelated." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Router/load-balancing loss computation vs mm_token_type_ids defaulting; different model paths." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Video processor lookup vs Gemma4 device-placement mismatch; no overlap." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "Generic placeholder title vs FP8 inference on B200; no common bug." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json vs Mistral regex patch crash; distinct tokenizer bugs." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different config classes and attributes: Phi4Multimodal defaults vs Llama4 pad_token_id." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error vs init_empty_weights constructor-argument mismatch." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "torch flex_attention compilation bug vs missing TensorFlow backend dependency." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer AttributeError on tasks vs D-FINE loss behavior; unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer backend patch crash vs list_repo_files revision forwarding; different code paths." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "FlashAttention/Mistral load failure vs Kosmos2.5 OCR index error; unrelated." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Adafactor OOM in DeepSpeed vs optimizer.step being called twice; not the same defect." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "auto_docstring annotation parsing vs Gemma4 processor import dependency." + }, + { + "left": "issue:44336", + "right": "issue:46851", + "accept": false, + "reason": "ANSI escape cleanup vs variable-length position_ids handling." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized model saving vs dequantized model loading with offload; different failures." + }, + { + "left": "issue:45859", + "right": "issue:46851", + "accept": false, + "reason": "CPU-offload rotary embedding incompatibility vs variable-length input handling." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Missing files in processor.save_pretrained vs a performance regression in data filtering." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "MLA q_a_layernorm epsilon config vs torch.split return values in GlmMoeDsaIndexer." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "auto_mappings stale config references vs leftover dummy classes in dummy_pt_objects." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Missing TensorFlow in extras install vs Python version support regression." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT masked-LM predictions vs DeepSeek tokenizer output regression." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "processor.save_pretrained issue vs ZeRO-3 from_pretrained; different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 42, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:41628", + "issue:42491", + "issue:42796", + "issue:43299", + "issue:43441", + "issue:43525", + "issue:43526", + "issue:43540", + "issue:43643", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44188", + "issue:44246", + "issue:44297", + "issue:44336", + "issue:44373", + "issue:44496", + "issue:44589", + "issue:45003", + "issue:45072", + "issue:45083", + "issue:45103", + "issue:45325", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45800", + "issue:45823", + "issue:46291", + "issue:46399", + "issue:46614", + "issue:46655", + "issue:46665", + "issue:46726", + "issue:46752", + "issue:46829", + "issue:46851" + ], + "soft_pairs": [ + "issue:43673|issue:43761", + "issue:43901|issue:43994", + "issue:43901|issue:44246", + "issue:43723|issue:44162", + "issue:43824|issue:43976", + "issue:42796|issue:43441", + "issue:43525|issue:43881", + "issue:46291|issue:46399", + "issue:45542|issue:45800", + "issue:44188|issue:44336", + "issue:44016|issue:45823", + "issue:43525|issue:43643", + "issue:44589|issue:45362", + "issue:44336|issue:44373", + "issue:45325|issue:46851", + "issue:46655|issue:46726", + "issue:41628|issue:43976", + "issue:44246|issue:44496", + "issue:45083|issue:46851", + "issue:43784|issue:45542", + "issue:44297|issue:45397", + "issue:46399|issue:46851", + "issue:45561|issue:45563", + "issue:45103|issue:46399", + "issue:43526|issue:45003", + "issue:43540|issue:46752", + "issue:45072|issue:46829", + "issue:46614|issue:46665", + "issue:42491|issue:43299" + ], + "prepared_review_unit_hash": "7c1983e0a978bb23655a1b7acc358044bf13ae639202c8bc529ed669f45fa56b", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36457, + "estimated_input_tokens": 9115, + "estimated_eval_tokens": 18486 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "I found no safe duplicate merges in the soft pairs. Most matches are superficial keyword overlaps across different models, APIs, or regression types, not the same underlying bug.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue stands out: the cluster is heterogeneous, spanning unrelated subsystems and model families, so there is no clear duplicate target to anchor on.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43299 is the most concrete standalone regression in the set (a v5 loading breakage on the main branch), but it is still model-specific and does not represent the whole cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Different regressions: cache handling in chunked prefill vs CLIP vision hidden_states output." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch for return_all_scores vs incorrect SigLIP model/pipeline outputs." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Documentation drift vs intermittent import slowness; not the same bug." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading issue vs ESM2 model breakage; unrelated failure modes." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Missing imported model class vs Python version compatibility regression." + }, + { + "left": "issue:42796", + "right": "issue:43441", + "accept": false, + "reason": "Tokenizer BPE-marker handling vs FlashAttention failure in the same model family; different code paths." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Llama4Config attribute error vs GLM-4V model loading failure; different models and causes." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Kernel dependency/version check on import vs missing decorator in a Qwen3.5 kernel registration path." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "TensorBoard-without-TF backend error vs torch/transformers version incompatibility." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "torch.compile attention divergence vs ANSI escape output in loading report; unrelated." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error vs device-map dtype/device mismatch in Gemma4 forward." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing config attribute vs AutoConfig trust_remote_code field loss; different APIs." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Missing Float8 storage type vs Qwen3.5 chat crash; different code paths." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Terminal coloring bug vs incorrect docstring; not duplicates." + }, + { + "left": "issue:45325", + "right": "issue:46851", + "accept": false, + "reason": "Different Qwen model families and different position_ids bugs (rope scaling vs variable-length handling)." + }, + { + "left": "issue:46655", + "right": "issue:46726", + "accept": false, + "reason": "Video processor padding bug vs missing image processor/backend detection." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "Missing AutoImageProcessor export vs Python version support regression." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import slowdowns vs unrecognized model/config loading failure." + }, + { + "left": "issue:45083", + "right": "issue:46851", + "accept": false, + "reason": "Helper output-length calculation in qwen3_omni_moe vs variable-length position_ids in Qwen3.5." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Import-time NameError in sentence-transformers integration vs TensorFlow backend detection issue." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "tokenizer_config class naming mismatch vs zero3 from_pretrained failure." + }, + { + "left": "issue:46399", + "right": "issue:46851", + "accept": false, + "reason": "Kernel decorator registration crash vs position_ids variable-length bug; different subsystems." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "xdist race in testing utils vs stale warning in paged generate." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "auto_docstring crash under future annotations vs Qwen3.5 kernel registration error." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "ImageProcessorFast label reduction bug vs unsafe sys.modules access in modeling_utils." + }, + { + "left": "issue:43540", + "right": "issue:46752", + "accept": false, + "reason": "Qwen3OmniMoe video processing ValueError vs empty-list chat template IndexError." + }, + { + "left": "issue:45072", + "right": "issue:46829", + "accept": false, + "reason": "bf16 dtype mismatch in model inference vs incorrect hidden_states returned by VideoPrism classifier." + }, + { + "left": "issue:46614", + "right": "issue:46665", + "accept": false, + "reason": "Stale phonemizer backend state vs inconsistent from_pretrained behavior in classes; not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen3 MoE loading, but one is a LoRA compatibility issue and the other is a model-loading regression; too broad to merge safely." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 43, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:36246", + "issue:41950", + "issue:42757", + "issue:42886", + "issue:43257", + "issue:43262", + "issue:43295", + "issue:43475", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43901", + "issue:44162", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44351", + "issue:44451", + "issue:44479", + "issue:44485", + "issue:44496", + "issue:44749", + "issue:44779", + "issue:44871", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45292", + "issue:45362", + "issue:45375", + "issue:45397", + "issue:45561", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45600", + "issue:45759", + "issue:45859", + "issue:46132", + "issue:46369", + "issue:46399", + "issue:46506", + "issue:46655", + "issue:46752", + "issue:46821", + "issue:46851" + ], + "soft_pairs": [ + "issue:43531|issue:45362", + "issue:44246|issue:44871", + "issue:45588|issue:46851", + "issue:44295|issue:45397", + "issue:44242|issue:45593", + "issue:44336|issue:44485", + "issue:45003|issue:46821", + "issue:42886|issue:46132", + "issue:43550|issue:46506", + "issue:44451|issue:46752", + "issue:44779|issue:45292", + "issue:36246|issue:42757", + "issue:43901|issue:45003", + "issue:43257|issue:45137", + "issue:36246|issue:44351", + "issue:41950|issue:44220", + "issue:45003|issue:45561", + "issue:44496|issue:45003", + "issue:43540|issue:46655", + "issue:45375|issue:45759", + "issue:43262|issue:44246", + "issue:44162|issue:45200", + "issue:44479|issue:46655", + "issue:45859|issue:46399", + "issue:44295|issue:44749", + "issue:43475|issue:46369", + "issue:45137|issue:45600", + "issue:43295|issue:45292", + "issue:45292|issue:45584" + ], + "prepared_review_unit_hash": "f881494e65a164bc7d58c7f2ee8716a9764ad41e64669ba831279ede070770a4", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36306, + "estimated_input_tokens": 9077, + "estimated_eval_tokens": 18410 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated Transformers bugs with only superficial overlap in model families or subsystems. I did not find any soft pair that clearly describes the same underlying bug/change, so all soft edges are rejected.", + "confidence": 0.94, + "canonical_issue_reason": "Issue 44479 is the broadest and most representative regression here: it spans multiple Qwen multimodal/video variants and has a concrete version-specific failure mode, making it the best cluster anchor among these issues.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44479 is the best representative issue because it has the widest blast radius and clearest regression framing, which makes it the most useful triage anchor in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Different bugs: Qwen3-MoE sliding_window behavior vs a Qwen3.5 chat crash. Same family, different code paths and symptoms." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Unrelated: a slow import problem versus a Gemma eos_token_id config mismatch." + }, + { + "left": "issue:45588", + "right": "issue:46851", + "accept": false, + "reason": "Both involve generation/input handling, but one is a flash-attention sinkless-model crash and the other is variable-length position_ids handling; not the same failure." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Different symptoms and areas: buffer registration access error versus a Gemma-4 ZeRO3 from_pretrained issue." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Both mention missing losses, but they affect different models and mechanisms; too broad to treat as one bug." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI/terminal output bug versus GLM-5 RoPE implementation; no shared underlying code-path." + }, + { + "left": "issue:45003", + "right": "issue:46821", + "accept": false, + "reason": "Modeling_utils sys.modules access bug is unrelated to the TimesFM window_size AttributeError." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline tokenizer cache loading and AttentionInterface.register behavior change are different problems." + }, + { + "left": "issue:43550", + "right": "issue:46506", + "accept": false, + "reason": "Different runtime failures in different models/hardware paths; no evidence of the same bug." + }, + { + "left": "issue:44451", + "right": "issue:46752", + "accept": false, + "reason": "Loading a specific model from cache is unrelated to apply_chat_template crashing on an empty list." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer output regression and resize_token_embeddings not updating output embeddings are distinct issues." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols, packages, and contexts." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs/behavior mismatch in TextClassificationPipeline is unrelated to unsafe sys.modules access." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both mention DeepSpeed, but one is about Qwen3 MOE weight conversion and the other is a ZeRO3 deque error." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Import errors for different APIs; not the same missing-export bug." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video pipeline choosing image processors and _torch_extract_fbank_features() are different audio/video issues." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "sys.modules access in modeling_utils is unrelated to pytest-xdist capture-file races." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading is a different failure mode than unsafe sys.modules access." + }, + { + "left": "issue:43540", + "right": "issue:46655", + "accept": false, + "reason": "Both are video-related, but they affect different models and the described causes differ (video input validation vs padding calculation)." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Both concern Qwen multimodal config handling, but one is a missing config field and the other is AutoModelForCausalLM.from_config unwrapping text_config; not the same bug." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat-template sampling-rate behavior is unrelated to import performance." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma-4 mm_token_type_ids defaults are unrelated model-specific bugs." + }, + { + "left": "issue:44479", + "right": "issue:46655", + "accept": false, + "reason": "Both mention video regressions, but they are for different model families and likely different processor logic." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "Related to Qwen3.5 rotary embedding code, but the reported failures are different enough that they do not look like one concrete bug." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "An attribute error around position_ids is unrelated to a performance regression after upgrading Transformers." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "SAM 3 Video encoder attribute loss and SAM3LiteText memory overflow are different model-specific failures." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque failure and broken auto_mappings references are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Processor.tokenizer regression and resize_token_embeddings/output_embeddings mismatch are different API bugs." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resizing and Whisper empty-transcription handling have no shared underlying code-path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 44, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:29870", + "issue:30333", + "issue:36246", + "issue:42199", + "issue:42886", + "issue:43065", + "issue:43352", + "issue:43408", + "issue:43475", + "issue:43526", + "issue:43540", + "issue:43596", + "issue:43653", + "issue:43723", + "issue:44038", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44368", + "issue:44466", + "issue:44479", + "issue:44623", + "issue:44857", + "issue:44877", + "issue:44977", + "issue:44987", + "issue:45229", + "issue:45250", + "issue:45335", + "issue:45375", + "issue:45397", + "issue:45412", + "issue:45459", + "issue:45563", + "issue:45656", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46170", + "issue:46506", + "issue:46710", + "issue:46726", + "issue:46832", + "issue:46851", + "issue:46894", + "issue:46897" + ], + "soft_pairs": [ + "issue:44877|issue:45375", + "issue:44466|issue:45335", + "issue:36246|issue:43596", + "issue:44242|issue:44857", + "issue:44623|issue:46726", + "issue:43526|issue:46897", + "issue:44987|issue:45459", + "issue:44038|issue:46506", + "issue:29870|issue:30333", + "issue:42199|issue:46851", + "issue:44060|issue:44077", + "issue:45412|issue:46123", + "issue:44479|issue:46170", + "issue:43526|issue:46894", + "issue:43653|issue:44162", + "issue:44368|issue:44977", + "issue:43352|issue:45250", + "issue:42886|issue:45229", + "issue:44877|issue:46170", + "issue:43475|issue:43540", + "issue:45397|issue:45656", + "issue:46170|issue:46832", + "issue:43065|issue:43408", + "issue:43723|issue:46710", + "issue:44368|issue:45563", + "issue:45987|issue:46082" + ], + "prepared_review_unit_hash": "f4142b2fbaf59ca89e479ac8a40e8bc90062fba75532495bc75a8d9d4c6aaae9", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35073, + "estimated_input_tokens": 8769, + "estimated_eval_tokens": 17794 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is not a single duplicate cluster; it contains many unrelated issues across tokenizers, vision/video models, config loading, training, and runtime behavior. The soft links are mostly superficial thematic matches, not the same underlying bug.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this heterogeneous set; the items describe distinct bugs in different subsystems and model families.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue: none of these is representative of the whole set, and several only share broad surface symptoms rather than a common code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both involve strict config handling, but one is about loading granite_speech and the other about a missing Qwen3.5MoeVisionConfig field; different model/config paths." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both touch tied/embedding weights, but one is inconsistent lm_head serialization by device and the other is resize_token_embeddings not updating decoder.embed_tokens." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Completely different failures: missing Qwen2.5 VL image processor import vs DeepSpeed ZeRO-3/BertModel index error." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Different bugs in different code paths: MoE load-balancing loss omission vs RT-DETR loss crash under fp16/CUDA." + }, + { + "left": "issue:44623", + "right": "issue:46726", + "accept": false, + "reason": "Related to processor/loading, but one is save_pretrained missing files and the other is AutoProcessor backend-missing error on load." + }, + { + "left": "issue:43526", + "right": "issue:46897", + "accept": false, + "reason": "Same broad label of training/processing bugs, but different models and likely different fixes; not the same underlying issue." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "One is a model-loading regression in v5.1+, the other is tokenizer error masking when protobuf is absent; unrelated root causes." + }, + { + "left": "issue:44038", + "right": "issue:46506", + "accept": false, + "reason": "Different subsystems and symptoms: Qwen3-VL-Moe compatibility bug vs PPDocLayoutV3 CPU/CUDA result mismatch." + }, + { + "left": "issue:29870", + "right": "issue:30333", + "accept": false, + "reason": "UnboundLocalError during generation for BART vs MLflow reporting not updated after failed training job; unrelated." + }, + { + "left": "issue:42199", + "right": "issue:46851", + "accept": false, + "reason": "DETR cardinality/background-class logic vs Qwen3.5 variable-length position_ids handling; different model families and code paths." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Both are warning/config-related, but one is a tied-weights warning and the other is an invalid optional post_init allowance; not the same bug." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "Memory-leak-on-delete in RT-DETR is unrelated to MaskGenerationPipeline's final partial batch handling." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Qwen video-input regression is unrelated to Windows/ROCm import and torch.distributed dependency problems." + }, + { + "left": "issue:43526", + "right": "issue:46894", + "accept": false, + "reason": "Both mention a training-loss shift pattern, but they are different model-specific reports and not enough evidence of one shared bug." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token decode bug vs ESM2 being broken; different components and failure modes." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "Both mention Qwen3.5, but one is a tied-embeddings warning and the other is a flash-attention generation regression." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "Generic Flash Attention 2.0 topic vs NemotronHForCausalLM not supporting FA2; same feature area, different concrete issue." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Tokenizer offline cache loading and Gemma4 multi-GPU CUDA OOM are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Config strictness for granite_speech is unrelated to Windows/ROCm import failures." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "SAM 3 video attribute error vs Qwen3OmniMoe video-input ValueError; different models and code paths." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "gemma-4 ZeRO-3 from_pretrained issue is unrelated to the deepspeed optimizer-step double-call report." + }, + { + "left": "issue:46170", + "right": "issue:46832", + "accept": false, + "reason": "Windows ROCm/torch.distributed import problem vs RT-DETR forward feature-level mismatch; not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "SAM3 pixel decoder dummy Conv2d vs sam3_video/sam3_tracker config mismatch; related family, different issue." + }, + { + "left": "issue:43723", + "right": "issue:46710", + "accept": false, + "reason": "Tokenizer loading issue in v5 vs DeepSeek-R1 accuracy regression after upgrading Transformers; different symptoms and likely causes." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "Qwen3.5 tied-weights warning is unrelated to paged generate's stale num_return_sequences warning." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache get_seq_length return-type bug vs LlamaConfig head_dim validation; no shared underlying defect." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 45, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:41628", + "issue:42994", + "issue:43065", + "issue:43232", + "issue:43441", + "issue:43475", + "issue:43482", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43701", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44483", + "issue:44623", + "issue:44749", + "issue:44869", + "issue:44871", + "issue:44933", + "issue:44964", + "issue:44995", + "issue:45103", + "issue:45137", + "issue:45200", + "issue:45356", + "issue:45375", + "issue:45478", + "issue:45588", + "issue:45656", + "issue:45685", + "issue:46123", + "issue:46132", + "issue:46202", + "issue:46421", + "issue:46431", + "issue:46726", + "issue:46752", + "issue:46821", + "issue:46851", + "issue:46858" + ], + "soft_pairs": [ + "issue:44995|issue:45200", + "issue:43931|issue:44368", + "issue:44188|issue:46202", + "issue:44483|issue:46202", + "issue:43906|issue:46202", + "issue:44188|issue:46421", + "issue:44995|issue:45588", + "issue:45685|issue:46431", + "issue:45103|issue:46858", + "issue:43065|issue:46123", + "issue:44361|issue:44749", + "issue:44297|issue:44623", + "issue:43994|issue:46726", + "issue:44749|issue:45656", + "issue:43701|issue:44749", + "issue:44871|issue:46132", + "issue:43646|issue:44162", + "issue:44623|issue:45656", + "issue:43441|issue:46752", + "issue:43475|issue:43643", + "issue:42994|issue:43482", + "issue:45375|issue:45478", + "issue:44188|issue:46851", + "issue:28282|issue:30990", + "issue:43475|issue:46821", + "issue:43232|issue:43673", + "issue:44869|issue:45356", + "issue:44964|issue:45137", + "issue:41628|issue:44933" + ], + "prepared_review_unit_hash": "3d4fd37f9d29390dd5d7a3fe77dc1c5edc2f327369dfa29c6fe60e1e5e5b04f9", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36459, + "estimated_input_tokens": 9115, + "estimated_eval_tokens": 18486 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers regressions. Only one soft pair looks plausibly like the same underlying bug: Qwen3.5 MoE config fields being dropped, which can surface as a from_pretrained failure. The rest are symptom-level similarities across different models/code paths and should stay separate.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45375 is the clearest root-cause report in the only plausible duplicate pair: it names the missing Qwen3_5MoeVisionConfig field directly, which is more canonical than the broader from_pretrained error in issue:45478.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45375 is the best anchor because it describes a concrete, fixable configuration-field omission rather than a downstream symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Different failures: stale indexer cache on second forward pass vs missing mm_token_type_ids defaulting for text-only fine-tuning." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Both involve Qwen, but one is a shape mismatch on load and the other is a tie_word_embeddings warning during LoRA training; not the same bug." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "Different code paths: torch.compile attention-kernel divergence vs Gemma pipeline roundtrip/tokenization issue." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "API request rejection in chat completions is unrelated to Gemma pipeline roundtrip behavior." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "A reproduction of another issue is not the same as the Gemma pipeline roundtrip failure." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Both mention cache/compile behavior, but one is attention-kernel divergence and the other is spurious recompilation from lazy cache init; too different to merge." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Second-forward stale cache bug is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MPS histogram/int dtype error is unrelated to missing chat template handling." + }, + { + "left": "issue:45103", + "right": "issue:46858", + "accept": false, + "reason": "Auto-docstring annotation parsing crash is a different subsystem from cache.get_seq_length() returning a device tensor in prefill." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "Dummy Conv2d placeholder in Sam3PixelDecoder is unrelated to MaskGenerationPipeline dropping final partial batch results." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "Tokenizer AttributeError and a Chinese performance regression are not the same underlying bug." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save metadata mismatch and processor save missing files are both serialization-related, but not the same concrete failure." + }, + { + "left": "issue:43994", + "right": "issue:46726", + "accept": false, + "reason": "Wrong outputs from AutoModel/pipeline are unrelated to missing vision-backend handling in AutoProcessor." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Performance regression during filtering is unrelated to DeepSpeed causing optimizer.step() to run twice." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch is a different problem from the reported data filtering slowdown." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id inconsistency is unrelated to AttentionInterface.register changing registered-function behavior." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Both are v5 breakages, but one is custom model initialization and the other is ESM2 breakage; too broad to treat as the same bug." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Processor save_pretrained missing files is unrelated to duplicated optimizer steps under DeepSpeed." + }, + { + "left": "issue:43441", + "right": "issue:46752", + "accept": false, + "reason": "FlashAttention failure for Ministral-3 is unrelated to apply_chat_template crashing on an empty list." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "Missing fpn_position_embeddings in SAM 3 video is not the same as trust_remote_code causing missing fields in AutoConfig output." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure and GGUF loading failure are opposite directions and different code paths." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": true, + "reason": "These appear to describe the same Qwen3.5 MoE config bug: a missing deepstack_visual_indexes field being dropped by strict validation, which then surfaces as a from_pretrained error in newer transformers." + }, + { + "left": "issue:44188", + "right": "issue:46851", + "accept": false, + "reason": "Attention-kernel branching under torch.compile is unrelated to Qwen3.5 variable-length position_ids handling." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch import error for AutoModel is unrelated to Sentence Transformers getting stuck while loading." + }, + { + "left": "issue:43475", + "right": "issue:46821", + "accept": false, + "reason": "SAM 3 video missing an attribute is unrelated to TimesFM window_size raising AttributeError." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both touch generation internals, but sync_gpus/update_model_kwargs and chunked_prefill cache initialization are distinct bugs." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decode crash and Kimi tokenizer codec regression are different tokenizer/audio issues." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Phi-4 multimodal loading failure is unrelated to the DeepSpeed ZeRO3 deque IndexError." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Missing AutoImageProcessor import and a nonexistent image_utils import are similar in theme but not the same bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet: it stays at a high level, distinguishes the one plausible Qwen3.5 MoE config duplicate from the unrelated regressions, and the soft-edge verdicts are generally conservative. The accepted 45375/45478 pair is reasonably supported by the titles and is phrased cautiously enough." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 46, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:36010", + "issue:39692", + "issue:41628", + "issue:42175", + "issue:42371", + "issue:42673", + "issue:42831", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43504", + "issue:43525", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:43986", + "issue:44038", + "issue:44222", + "issue:44351", + "issue:44373", + "issue:44623", + "issue:44792", + "issue:44857", + "issue:44933", + "issue:44945", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45230", + "issue:45310", + "issue:45375", + "issue:45397", + "issue:45406", + "issue:45561", + "issue:45593", + "issue:45600", + "issue:45753", + "issue:46202", + "issue:46326", + "issue:46399", + "issue:46463", + "issue:46701", + "issue:46726", + "issue:46851", + "issue:46877" + ], + "soft_pairs": [ + "issue:45310|issue:45375", + "issue:46463|issue:46877", + "issue:39692|issue:46726", + "issue:44373|issue:46851", + "issue:45230|issue:45397", + "issue:43986|issue:46726", + "issue:44038|issue:44623", + "issue:43723|issue:46326", + "issue:42994|issue:43525", + "issue:42371|issue:45092", + "issue:43873|issue:44945", + "issue:45406|issue:46202", + "issue:43784|issue:46399", + "issue:44945|issue:45753", + "issue:44857|issue:46877", + "issue:45593|issue:46877", + "issue:41628|issue:44351", + "issue:36010|issue:41628", + "issue:45081|issue:45103", + "issue:46202|issue:46701", + "issue:43278|issue:45137", + "issue:44038|issue:45310", + "issue:42673|issue:43704", + "issue:44222|issue:45310", + "issue:42994|issue:43824", + "issue:42831|issue:44222", + "issue:43504|issue:45561", + "issue:43232|issue:44792", + "issue:45561|issue:45600", + "issue:42175|issue:44933" + ], + "prepared_review_unit_hash": "aad4afa711e752f46473039376d29b0b50d7ffdbdf799bca00b8bc0ae9d3e70c", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36309, + "estimated_input_tokens": 9078, + "estimated_eval_tokens": 18412 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a heterogeneous set of Transformers bug reports spanning unrelated subsystems. The only clear duplicate pair is the Qwen3ForCausalLM VRAM leak report (42673/43704); the rest are too loosely related to merge safely.", + "confidence": 0.93, + "canonical_issue_reason": "issue:42673 is the more complete report for the exact duplicate VRAM leak/threading bug and has substantially more discussion than issue:43704.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the best canonical issue: it precisely describes the bug, has richer context, and cleanly subsumes issue:43704.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Same model family, but one is a from_pretrained error and the other is a missing config field; not clearly the same concrete bug." + }, + { + "left": "issue:46463", + "right": "issue:46877", + "accept": false, + "reason": "Unrelated problems: T5 decoder-input validation messaging vs auto_docstring stdout diagnostics." + }, + { + "left": "issue:39692", + "right": "issue:46726", + "accept": false, + "reason": "Both involve processor/vision docs or backend handling, but they are different failure modes and code paths." + }, + { + "left": "issue:44373", + "right": "issue:46851", + "accept": false, + "reason": "Both mention position_ids, but one is a docstring issue and the other is a runtime bug in Qwen3.5 variable-length handling." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic placeholder bug report versus a specific Gemma-4 ZeRO3 regression; no evidence they are the same issue." + }, + { + "left": "issue:43986", + "right": "issue:46726", + "accept": false, + "reason": "Both concern missing vision backends, but one is a video-model AutoProcessor crash and the other is an unrecognized image-processor error." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Different bugs: Qwen3-VL-Moe loading versus processor.save_pretrained missing files." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer loading failure in v5 and a mistral_common continue_final_message bug are separate problems." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized model saving failure and missing pad_token_id on Llama4Config are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 API guidance and InternVL2 meta-initialization incompatibility are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization offloading behavior and pipeline-parallel output correctness are different code paths." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Gemma processor attribute crash and Gemma pipeline roundtrip failure are distinct bugs." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Import-time NameError in sentence-transformers integration versus a Qwen3.5 kernel registration crash." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallel output corruption and a Qwen3_5 image-specific infinite loop do not match." + }, + { + "left": "issue:44857", + "right": "issue:46877", + "accept": false, + "reason": "CUDA AMP loss crash and auto_docstring import-time logging are unrelated." + }, + { + "left": "issue:45593", + "right": "issue:46877", + "accept": false, + "reason": "D-FINE auxiliary-loss behavior and docstring diagnostics are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they are for different symbols with different likely fixes." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Different missing exports/imports from different parts of transformers; not the same bug." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Two separate helper-function crashes with different triggers and stack paths." + }, + { + "left": "issue:46202", + "right": "issue:46701", + "accept": false, + "reason": "Both mention Gemma, but one is a tokenizer roundtrip issue and the other is an image-text-to-text pipeline failure." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift in eval and a DeepSpeed ZeRO3 deque IndexError are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Related model family, but the reported failures differ: one is a Qwen3-VL-Moe bug and the other is a Qwen3.5 MoE from_pretrained regression." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same VRAM leak report for Qwen3ForCausalLM used across multiple dataloader threads; 43704 is a duplicate of 42673." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained failure and Qwen3.5 Moe from_pretrained loading error are different concrete problems." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized save failure and missing Qwen2_5_VL export are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both are FP8-related, but one is an accuracy regression and the other is a save_pretrained bug." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Beit legacy-field loading bug and pytest-xdist race on captured_info.txt are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and janus image generation test failure are different issues." + }, + { + "left": "issue:45561", + "right": "issue:45600", + "accept": false, + "reason": "Testing-utils xdist race and broken auto_mappings references are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Backend packaging omission for TensorFlow and a missing image_utils import are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The canonical choice and cluster summary are grounded in the packet: 42673 and 43704 have identical titles and the discussion counts support choosing 42673 as canonical. The soft-edge verdicts are also conservative overall, with no strong overclaims beyond what the titles support." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 47, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:27957", + "issue:30333", + "issue:30990", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:39692", + "issue:42175", + "issue:42222", + "issue:42673", + "issue:42898", + "issue:43116", + "issue:43295", + "issue:43479", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44016", + "issue:44079", + "issue:44190", + "issue:44295", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44625", + "issue:44734", + "issue:44779", + "issue:44869", + "issue:44933", + "issue:45092", + "issue:45103", + "issue:45200", + "issue:45310", + "issue:45356", + "issue:45405", + "issue:45561", + "issue:45656", + "issue:45712", + "issue:45812", + "issue:46077", + "issue:46123", + "issue:46231", + "issue:46693", + "issue:46829", + "issue:46877", + "issue:46909" + ], + "soft_pairs": [ + "issue:44496|issue:45561", + "issue:43976|issue:44933", + "issue:45405|issue:46877", + "issue:30333|issue:33290", + "issue:30990|issue:35141", + "issue:42898|issue:44568", + "issue:42175|issue:45712", + "issue:43479|issue:46829", + "issue:44623|issue:45310", + "issue:46077|issue:46123", + "issue:42673|issue:46693", + "issue:43479|issue:44869", + "issue:42898|issue:45812", + "issue:43723|issue:44568", + "issue:44016|issue:46231", + "issue:44625|issue:46877", + "issue:39692|issue:46909", + "issue:42898|issue:44779", + "issue:43479|issue:44079", + "issue:43784|issue:43824", + "issue:44734|issue:45103", + "issue:43874|issue:45092", + "issue:27957|issue:30333", + "issue:42898|issue:43295", + "issue:42222|issue:45200", + "issue:30333|issue:34689", + "issue:42898|issue:45356", + "issue:44295|issue:45656", + "issue:43116|issue:44190", + "issue:45405|issue:45712" + ], + "prepared_review_unit_hash": "a98a17444645501614cb7842ef1b576711da33c96c324652719f3afe1e003849", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36964, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated Transformers issues with only superficial similarity around regressions, tokenization, multimodal loading, and tooling. I do not see any soft pair that looks like the same underlying bug/change.", + "confidence": 0.88, + "canonical_issue_reason": "If a single anchor is needed, issue 45812 is the strongest representative: it is a concrete, high-severity silent regression affecting all Granite tokenizers.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45812 is the most actionable and broadly impactful report in the set, with a clear user-facing correctness bug and reproducible symptoms.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Different subsystems and failure modes: model loading/config resolution vs pytest-xdist file-race behavior." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Unrelated problems: Python compatibility claim vs a bad import path in image_utils." + }, + { + "left": "issue:45405", + "right": "issue:46877", + "accept": false, + "reason": "Release/versioning complaint vs import-time docstring lint output; no shared bug." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow reporting after failure and DeepSpeed optimizer OOM are distinct training issues." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Loading hang in sentence-transformers does not match embedding reinitialization in post_init." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer-related, but one is v5 cleanup-space behavior and the other is special-token insertion for a specific tokenizer." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Missing TensorFlow in an install extra is unrelated to leftover dummy PT classes." + }, + { + "left": "issue:43479", + "right": "issue:46829", + "accept": false, + "reason": "Different model families and bugs: default config initialization vs wrong output field assignment." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Saving processor artifacts missing files is not the same as a Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Attention-mask construction for Gemma4 is unrelated to final-partial-batch handling in MaskGenerationPipeline." + }, + { + "left": "issue:42673", + "right": "issue:46693", + "accept": false, + "reason": "A VRAM leak in multithreaded dataloading is not the same as a long-context FA2 performance regression." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Phi4 multimodal config defaults and Whisper timestamp decode crash are different code paths." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both affect tokenization behavior in v5, but one is cleanup-space semantics and the other is wrong token IDs for Granite models." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading in v5 and add_special_tokens behavior for mdeberta-v3 are separate regressions." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error and documentation parameter-name mismatch are unrelated doc issues." + }, + { + "left": "issue:44625", + "right": "issue:46877", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 has nothing to do with auto_docstring logging diagnostics." + }, + { + "left": "issue:39692", + "right": "issue:46909", + "accept": false, + "reason": "Two different broken documentation examples with different models and failures." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but the affected models and failure mechanisms differ." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config defaulting in Phi4Multimodal and ModelOutput key assignment are unrelated internals." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Import-time NameError in sentence-transformers is not the same as a missing Qwen2.5-VL export." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "Serve KV-cache indexing crash and auto_docstring future-annotations crash are separate code paths." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image-patch method in GLM46V and InternVL2 meta-init incompatibility are different multimodal bugs." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention support for XLM-Roberta and MLflow reporting after training failure are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Tokenization-space behavior and processor/tokenizer/image regression in a custom model are not the same bug." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Broken VitPose model files and Gemma4 mm_token_type_ids defaults are separate multimodal issues." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLflow failure reporting and Llama 3.2 Vision model-loading break are unrelated regressions." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Different tokenizer regressions affecting different models and codecs." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position_ids buffer access error and duplicate optimizer-step behavior with DeepSpeed are distinct problems." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Empty multi-label classification results and local dataset loading failure are different example-script issues." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "Unreleased dependency pin and leftover dummy class cleanup are unrelated maintenance issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 48, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:30333", + "issue:30990", + "issue:36331", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42796", + "issue:42831", + "issue:43122", + "issue:43232", + "issue:43475", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:44060", + "issue:44075", + "issue:44190", + "issue:44222", + "issue:44263", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44448", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44933", + "issue:44977", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45245", + "issue:45310", + "issue:45656", + "issue:45712", + "issue:45834", + "issue:46082", + "issue:46132", + "issue:46209", + "issue:46231", + "issue:46305", + "issue:46424", + "issue:46614", + "issue:46710", + "issue:46762", + "issue:46932" + ], + "soft_pairs": [ + "issue:30333|issue:36331", + "issue:22355|issue:28282", + "issue:44222|issue:44297", + "issue:44075|issue:46231", + "issue:44060|issue:45125", + "issue:44448|issue:46132", + "issue:44933|issue:45712", + "issue:42796|issue:43723", + "issue:44360|issue:46762", + "issue:42831|issue:44623", + "issue:44297|issue:45656", + "issue:45245|issue:46132", + "issue:43479|issue:43644", + "issue:44222|issue:44623", + "issue:46132|issue:46710", + "issue:44373|issue:44485", + "issue:41628|issue:42757", + "issue:44297|issue:45310", + "issue:45071|issue:46082", + "issue:43723|issue:44977", + "issue:46614|issue:46932", + "issue:44222|issue:44484", + "issue:43122|issue:46209", + "issue:44190|issue:46082", + "issue:44263|issue:45092", + "issue:43232|issue:46424", + "issue:43232|issue:45834", + "issue:43475|issue:46305", + "issue:30990|issue:38175" + ], + "prepared_review_unit_hash": "faa3ed36ba6d8733ba2a09f561681f046d9d5ff142bf502711e3750649e06cb1", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36037, + "estimated_input_tokens": 9010, + "estimated_eval_tokens": 18276 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues; the soft pairs are only superficially similar and do not describe the same underlying bug/change. No PRs are present.", + "confidence": 0.43, + "canonical_issue_reason": "No single issue is a clear canonical duplicate target because the set spans many unrelated regressions, import errors, tokenizer/model bugs, and version-migration problems.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45071 is the best representative only in the loose sense that it is a broad Transformers v5 regression report; however, it is not a true duplicate of the other issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "Different failures: MLflow run-state reporting after training job failure vs. a CustomTrainer method signature error." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Both are import errors, but for different missing modules/libraries and different root causes." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "Both touch FP8/tokenizer save flows, but one is about MOE FP8 save_pretrained while the other is a tokenizer_config class mismatch." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Optimizer arguments not being used is unrelated to a docs typo about set_training parameter naming." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Both are Qwen3-family MoE issues, but one is an incorrect tied-weights warning and the other is missing tensor-parallel planning." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Different problems: output regression across versions vs. AttentionInterface.register changing registered-function behavior." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Both concern missing/nonexistent imports, but they affect different modules and different cleanup issues." + }, + { + "left": "issue:42796", + "right": "issue:43723", + "accept": false, + "reason": "Tokenizer BPE-marker handling and AutoTokenizer loading failures are distinct bugs." + }, + { + "left": "issue:44360", + "right": "issue:46762", + "accept": false, + "reason": "Both mention numerical behavior in attention-like code, but they are different implementations with different observed problems." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FineGrainedFP8 accuracy and processor.save_pretrained missing files are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save-pretrained metadata mismatch is unrelated to optimizer step being called twice under deepspeed." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category-count runtime error is unrelated to AttentionInterface.register behavior changes." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Config default initialization bug and non-persistent buffer corruption are different code paths." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both involve saving artifacts, but one is FP8 MOE model persistence and the other is processor save file completeness." + }, + { + "left": "issue:46132", + "right": "issue:46710", + "accept": false, + "reason": "AttentionInterface.register semantics and an accuracy regression after upgrade are not the same bug." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring mismatch for position_ids is unrelated to GLM-5 RoPE implementation behavior." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Both are import-related, but they concern different symbols from different packages and likely different breakages." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save-pretrained metadata mismatch is not the same as Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "PretrainedConfig type-checking regression and LlamaConfig head_dim validation are different configuration bugs." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "AutoTokenizer loading failure and Qwen3.5 flash-attention generation issues are unrelated." + }, + { + "left": "issue:46614", + "right": "issue:46932", + "accept": false, + "reason": "Both are tokenizer/processing-related multilingual issues, but one is stale phonemizer state and the other is Gemma4 quantized embedder dtype casting." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained moe and max_shard_size default sizing are different save-path concerns with different symptoms." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenization differences across versions and FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Local dataset loading in a training script is unrelated to LlamaConfig head_dim validation." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return values in a specific indexer are unrelated to remote-code checkpoint incompatibility with meta initialization." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and cache max length behavior are separate generation/cache bugs." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs update bug is unrelated to Kosmos2.5 OCR index error." + }, + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "SAM 3 Video missing attribute and MoonshineStreaming off-by-one mask logic are distinct model bugs." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang and Siglip2 zero-probabilities are different model-specific issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 49, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:30990", + "issue:39401", + "issue:41628", + "issue:42757", + "issue:42796", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43644", + "issue:43824", + "issue:44038", + "issue:44188", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44295", + "issue:44360", + "issue:44373", + "issue:44483", + "issue:44568", + "issue:44623", + "issue:44829", + "issue:45092", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46050", + "issue:46209", + "issue:46305", + "issue:46328", + "issue:46369", + "issue:46402", + "issue:46431", + "issue:46655", + "issue:46909" + ], + "soft_pairs": [ + "issue:46305|issue:46655", + "issue:46209|issue:46431", + "issue:44295|issue:44623", + "issue:44263|issue:44360", + "issue:44263|issue:44373", + "issue:45092|issue:45357", + "issue:42907|issue:46050", + "issue:46369|issue:46909", + "issue:42831|issue:45310", + "issue:45381|issue:46305", + "issue:44279|issue:46328", + "issue:44483|issue:45406", + "issue:22355|issue:27957", + "issue:43644|issue:46305", + "issue:44261|issue:45092", + "issue:44829|issue:46002", + "issue:42831|issue:44038", + "issue:43408|issue:46402", + "issue:41628|issue:43824", + "issue:42831|issue:45753", + "issue:22355|issue:28218", + "issue:39401|issue:44568", + "issue:42886|issue:43066", + "issue:27957|issue:30990", + "issue:42886|issue:42994", + "issue:44188|issue:44373", + "issue:42757|issue:45278", + "issue:43408|issue:46369", + "issue:43232|issue:45593", + "issue:42796|issue:43122" + ], + "prepared_review_unit_hash": "5928d1736752922c95533b2845d253d98e167b72528081179d72b44620613e94", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36190, + "estimated_input_tokens": 9048, + "estimated_eval_tokens": 18352 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft matches are mostly superficial title similarity across unrelated regressions (tokenizer, import, model-loading, quantization, and processor bugs). I would not merge any of these pairs as duplicates.", + "confidence": 0.81, + "canonical_issue_reason": "If a representative issue is needed, issue 43122 is the broadest tokenizer-regression report in the set (version-to-version behavior change), but it is only a loose thematic anchor, not a true duplicate umbrella.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43122 is the least-specific representative among the tokenizer-related items, so it is the best single anchor here despite the cluster being too heterogeneous for a real canonical duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46305", + "right": "issue:46655", + "accept": false, + "reason": "Different bugs: sliding-window attention off-by-one vs video padding calculation." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 inference failure on B200 is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Reading a buffer after registration and missing save_pretrained files are different code paths." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "torch.split return handling is not the same as a missing ReLU in the DSA indexer." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "A runtime indexing bug and a docstring typo are not duplicates." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Meta-init incompatibility for old InternVL2 and incorrect visual-encoder keys on save are different problems." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Both touch quantized save/load, but the quantization stacks and failure modes differ." + }, + { + "left": "issue:46369", + "right": "issue:46909", + "accept": false, + "reason": "Memory overflow in sam3-litetext is unrelated to a broken documentation example." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FP8 accuracy degradation and Qwen3.5 MoE from_pretrained failure are not the same bug." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Both are indexing/positioning issues, but they affect different models and different mechanisms." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "A dependency problem and a SequenceFeatureExtractor.pad performance issue are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request rejection is unrelated to Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "Missing transformers.onnx and XLMRoberta Flash Attention 2 are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer junk and sliding-window mask off-by-one are different failures." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "Missing rms_norm_eps precision issue is not the same as a meta-initialization incompatibility." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Both mention training instability, but the concrete causes and code paths differ." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FP8 accuracy regression and a Qwen3-VL-Moe bug are not the same underlying issue." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "A model-type warning and missing Sam3Model text encoder weights are different problems." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Missing AutoImageProcessor and missing Qwen2_5_VLForConditionalGeneration are separate import/export failures." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy issues and an infinite loop on a specific image are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Missing onnx module and extra tokenizer spacing after added tokens are different bugs." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Wrong offset_mapping and missing BOS/EOS insertion are distinct tokenizer behaviors." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure and wrong tokenizer decoder type are related only at a broad tokenizer level, not the same defect." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "Flash Attention 2 on XLMRoberta and Sentence Transformers loading stalls are unrelated." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Tokenizer offline-cache loading and quantized model saving are different code paths." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "torch.compile attention divergence and a position_ids docstring typo are not duplicates." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related regressions, but one is a specific huggingface_hub symbol import and the other is a broader upgrade-time import breakage." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "A type-mismatch warning is unrelated to a memory overflow in model execution." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs after sync_gpus and missing auxiliary losses are different features." + }, + { + "left": "issue:42796", + "right": "issue:43122", + "accept": false, + "reason": "Tokenizer BPE-marker handling for one model and general version-to-version tokenization drift are not the same concrete bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 50, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:22355", + "issue:29159", + "issue:30990", + "issue:33357", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42907", + "issue:43122", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43540", + "issue:43576", + "issue:43646", + "issue:43756", + "issue:43784", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44060", + "issue:44117", + "issue:44222", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44485", + "issue:44488", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45216", + "issue:45278", + "issue:45290", + "issue:45381", + "issue:45478", + "issue:45600", + "issue:46002", + "issue:46172", + "issue:46326", + "issue:46614", + "issue:46762", + "issue:46909" + ], + "soft_pairs": [ + "issue:43646|issue:46002", + "issue:45381|issue:46909", + "issue:43540|issue:46909", + "issue:30990|issue:33357", + "issue:42907|issue:44945", + "issue:42907|issue:44964", + "issue:39401|issue:44488", + "issue:43976|issue:44945", + "issue:44295|issue:44297", + "issue:43576|issue:43784", + "issue:42371|issue:44263", + "issue:43756|issue:44060", + "issue:46172|issue:46762", + "issue:42907|issue:43937", + "issue:41628|issue:45278", + "issue:43784|issue:45600", + "issue:44117|issue:44704", + "issue:39401|issue:44521", + "issue:44625|issue:44829", + "issue:43408|issue:43475", + "issue:44704|issue:45290", + "issue:43122|issue:44521", + "issue:44360|issue:44485", + "issue:43122|issue:46326", + "issue:42907|issue:43452", + "issue:44038|issue:44297", + "issue:22355|issue:29159", + "issue:45216|issue:45478", + "issue:44038|issue:44295", + "issue:39401|issue:46614", + "issue:44038|issue:44222" + ], + "prepared_review_unit_hash": "0124654e7956b7186de7f5b75008a097096d4b4106e0ad57be57e7d93f4cdf20", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 35975, + "estimated_input_tokens": 8994, + "estimated_eval_tokens": 18244 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are a set of mostly unrelated issues across different models, tokenizers, and release regressions. I found no soft-edge pair that clearly represents the same underlying bug/change, so all candidate links are rejected.", + "confidence": 0.94, + "canonical_issue_reason": "No strong duplicate cluster emerges; the issues span distinct code paths and model families, with only loose thematic overlap in a few cases.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the clearest standalone bug report: it has a specific, reproducible failure mode, broad impact for multimodal chat templating, and a focused fix target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Different problems: custom model init breakage vs NaNs during classification head checkout/init." + }, + { + "left": "issue:45381", + "right": "issue:46909", + "accept": false, + "reason": "Different subsystems: Qwen2.5-VL video position IDs vs a broken DIA documentation example." + }, + { + "left": "issue:43540", + "right": "issue:46909", + "accept": false, + "reason": "Qwen3OmniMoe video-input processing failure is unrelated to a documentation example bug." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Both are loading issues, but they affect different models and failure modes (stuck load vs bus error)." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals is unrelated to pipeline-parallelism output corruption." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Different model families and different workflows: dequantized save vs model load failure." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Wrong offset mappings in Qwen3 tokenizer is unrelated to loading cjvt/sleng-bert." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility regression is not the same bug as incorrect outputs under pipeline parallelism." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "One is a position_ids buffer access error; the other is a tokenizer_class mismatch on save." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken env command and a sentence-transformers import NameError are different failure paths." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 settings warning and a torch.split return-value issue are unrelated." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Both involve model internals, but Smollm3 RoPE-layer mismatch and Qwen3-Next tied-weights warning are distinct bugs." + }, + { + "left": "issue:46172", + "right": "issue:46762", + "accept": false, + "reason": "Different numerical issues in different models/attention implementations; not the same code-path problem." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Dequantized save failure is unrelated to invalid GenerationConfig handling." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "A single missing import is not the same as a broad post-upgrade import-regression report." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Sentence-transformers import failure and stale auto_mappings references are separate import-time issues." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Tokenizer mapping None-handling and AutoProcessor kwarg forwarding are different bugs." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer offset mapping bug is unrelated to multimodal assistant mask generation." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training degeneration are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Both are SAM3-related, but model-type warning and missing attribute error are distinct issues." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "cached_file kwargs forwarding and chat-template tool-call crashes are different code paths." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Tokenization differences across versions are not the same as multimodal assistant_mask failures." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer ReLU discussion and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Both concern tokenizer behavior, but they are separate bugs in different tokenizer stacks." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "GGUF loading breakage is unrelated to dequantized save failures." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe version bug and tokenizer_class save mismatch are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "No module import failure and slow/fast tokenizer inconsistency are different problems." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 and version regressions, but one is save_pretrained checkpoint correctness and the other is from_pretrained loading failure; not clearly the same concrete bug." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe bug and position_ids buffer access error do not share the same underlying failure." + }, + { + "left": "issue:39401", + "right": "issue:46614", + "accept": false, + "reason": "Qwen3 tokenizer offset mapping and stale phonemizer backend state are unrelated tokenizer issues." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Different Qwen3-VL-Moe issue vs FP8 save_pretrained moe bug; same broad area, different concrete failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 51, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:36296", + "issue:38617", + "issue:42175", + "issue:42371", + "issue:42673", + "issue:42757", + "issue:42898", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43493", + "issue:43611", + "issue:43618", + "issue:43824", + "issue:43825", + "issue:43874", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44075", + "issue:44188", + "issue:44351", + "issue:44355", + "issue:44485", + "issue:44610", + "issue:44625", + "issue:44779", + "issue:44936", + "issue:44938", + "issue:44945", + "issue:45072", + "issue:45083", + "issue:45200", + "issue:45278", + "issue:45357", + "issue:45446", + "issue:45478", + "issue:45753", + "issue:46002", + "issue:46231", + "issue:46328", + "issue:46614", + "issue:46829", + "issue:46877", + "issue:46909" + ], + "soft_pairs": [ + "issue:43824|issue:44351", + "issue:43116|issue:45200", + "issue:43122|issue:44779", + "issue:43278|issue:44945", + "issue:36296|issue:43116", + "issue:42757|issue:44351", + "issue:22355|issue:29127", + "issue:42175|issue:44938", + "issue:43493|issue:45753", + "issue:44188|issue:46328", + "issue:46231|issue:46877", + "issue:43937|issue:44485", + "issue:42898|issue:46614", + "issue:38617|issue:45278", + "issue:43066|issue:44779", + "issue:43611|issue:46002", + "issue:43976|issue:45446", + "issue:44038|issue:45478", + "issue:42673|issue:45072", + "issue:43116|issue:43493", + "issue:45357|issue:45478", + "issue:44355|issue:46231", + "issue:44625|issue:46002", + "issue:42371|issue:43874", + "issue:45083|issue:45446", + "issue:44610|issue:46909", + "issue:44075|issue:44936", + "issue:43618|issue:46829", + "issue:43116|issue:43825" + ], + "prepared_review_unit_hash": "33dc85a766262b8c5bb912822567482f77ee144f650f3d9823327f74eef09c73", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35984, + "estimated_input_tokens": 8996, + "estimated_eval_tokens": 18248 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most pairs share only broad vocabulary (v5 regressions, imports, tokenizers, multimodal models) but do not describe the same concrete bug or change. No soft edge looks safe to merge as a duplicate.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45278 is the broadest high-signal report in the set: a v5 upgrade causing multiple import failures. It is the closest thing to a hub for the cluster, though the cluster as a whole is not a true duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the best representative of the overall theme of upgrade/regression breakage. It is still too broad to unify the other items as duplicates, but it is the most central issue here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and likely different breakpoints; not the same bug." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Classification example output vs Gemma multimodal token ID handling are unrelated code paths." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they affect different models and symptoms; too broad to merge." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype drift during eval and pipeline-parallel incorrect outputs are different failure modes." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and example script multi-label empties are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Different import sources and missing names; same symptom class, not same underlying issue." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "One is a missing module import, the other is an error-message clarity complaint in LayoutLMv3." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Dependency/backend packaging issue vs Python 3.14 load failure; no shared concrete code path." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "Different multimodal model regressions with different behavior and triggers." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Attention-kernel divergence and feature-extractor padding performance are unrelated." + }, + { + "left": "issue:46231", + "right": "issue:46877", + "accept": false, + "reason": "Both touch docs, but one is a parameter-name doc bug and the other is import-time lint output." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM-5, but one is GenerationConfig validation and the other is RoPE implementation; different bugs." + }, + { + "left": "issue:42898", + "right": "issue:46614", + "accept": false, + "reason": "Tokenizer cleanup behavior and phonemizer backend caching are different subsystems." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Specific missing import vs broad post-upgrade import failures; 45278 is broader but not the same defect." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer-related regressions, but the concrete symptoms and affected paths differ." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Model loading with base_model_prefix and NaNs in classification heads are different problems." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python version compatibility failure and an AuxRequest version-check bug are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Same model family, but different versions and different loading failures; not one mergeable fix." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in multi-threaded dataloading is unrelated to bfloat16 dtype mismatches in CI." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example script classification behavior and SigLIP2 implementation discrepancy do not share a bug." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5/MoE regressions, but one is save_pretrained key handling and the other is from_pretrained loading." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled-file import errors and a documentation parameter mismatch are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "Config propagation for num_labels and NaNs on init/checkpoint are different failure modes." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 policy messaging and a missing image-patch method are unrelated." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Different helper/version-check bugs in different modules; no shared code path." + }, + { + "left": "issue:44610", + "right": "issue:46909", + "accept": false, + "reason": "Processor input-size mismatch and a broken documentation example for DIA are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer args not applied and trainer.evaluate() failing after train are separate training issues." + }, + { + "left": "issue:43618", + "right": "issue:46829", + "accept": false, + "reason": "Both involve output fields, but one is CLIP attentions missing and the other is VideoPrism hidden_states misassigned." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example script multi-label output bug and pipeline translation-task messaging are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 52, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28348", + "issue:33453", + "issue:39401", + "issue:42222", + "issue:42673", + "issue:42796", + "issue:43064", + "issue:43066", + "issue:43441", + "issue:43450", + "issue:43673", + "issue:43827", + "issue:43937", + "issue:43986", + "issue:43992", + "issue:44117", + "issue:44263", + "issue:44291", + "issue:44380", + "issue:44410", + "issue:44466", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44977", + "issue:44995", + "issue:45071", + "issue:45081", + "issue:45092", + "issue:45216", + "issue:45310", + "issue:45341", + "issue:45405", + "issue:45446", + "issue:45561", + "issue:45563", + "issue:45588", + "issue:46002", + "issue:46077", + "issue:46328", + "issue:46512", + "issue:46614", + "issue:46640", + "issue:46655", + "issue:46697", + "issue:46710", + "issue:46752", + "issue:46762" + ], + "soft_pairs": [ + "issue:43066|issue:46710", + "issue:43992|issue:44704", + "issue:46614|issue:46752", + "issue:44995|issue:46697", + "issue:44263|issue:46762", + "issue:45092|issue:46762", + "issue:28348|issue:46640", + "issue:43827|issue:45446", + "issue:33453|issue:39401", + "issue:43064|issue:43673", + "issue:42222|issue:43441", + "issue:42673|issue:42796", + "issue:43937|issue:45563", + "issue:43673|issue:46328", + "issue:44291|issue:45071", + "issue:42673|issue:46655", + "issue:42222|issue:46752", + "issue:45405|issue:45446", + "issue:44117|issue:45588", + "issue:44466|issue:46002", + "issue:44380|issue:46077", + "issue:39401|issue:44779", + "issue:43937|issue:44977", + "issue:44410|issue:45310", + "issue:45341|issue:45561", + "issue:45071|issue:46512", + "issue:45341|issue:46328", + "issue:43450|issue:43986", + "issue:44829|issue:45216", + "issue:44117|issue:45081" + ], + "prepared_review_unit_hash": "1130b9a068554166a69152e664ab1f9ecdb04df44b7c6a6dcc7a779a0aec9b29", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36667, + "estimated_input_tokens": 9167, + "estimated_eval_tokens": 18590 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is broad and mostly heterogeneous: tokenizer loading/format regressions, attention-backend issues, video/processor bugs, and a few training/generation failures. I don\u2019t see any soft pair that clearly describes the exact same underlying bug, so I would keep them separate. The most umbrella-like issue is the generic tokenizer-loading regression.", + "confidence": 0.57, + "canonical_issue_reason": "issue:33453 is the broadest, most umbrella-style report in the set (\u201cRegression in tokenizer loading\u201d) and best matches the cluster\u2019s repeated tokenizer-related regressions without overfitting to one model.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:33453 is the best representative issue because it is generic enough to cover several tokenizer-loading variants in the cluster, while the other issues are too model- or symptom-specific.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43066", + "right": "issue:46710", + "accept": false, + "reason": "Both are v5 regressions, but one is tokenizer decoder typing and the other is a model accuracy regression; different failures and code paths." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "UMT5 weight loading and AutoProcessor kwargs forwarding are unrelated loader issues." + }, + { + "left": "issue:46614", + "right": "issue:46752", + "accept": false, + "reason": "Stale phonemizer backend vs empty-list chat templating crash are different subsystems." + }, + { + "left": "issue:44995", + "right": "issue:46697", + "accept": false, + "reason": "GlmMoeDsa cache corruption and image over-padding are unrelated bugs." + }, + { + "left": "issue:44263", + "right": "issue:46762", + "accept": false, + "reason": "Different attention/model numerical issues; not the same concrete bug." + }, + { + "left": "issue:45092", + "right": "issue:46762", + "accept": false, + "reason": "Meta-init incompatibility vs attention numerical drift are distinct problems." + }, + { + "left": "issue:28348", + "right": "issue:46640", + "accept": false, + "reason": "A flash-attention support request and a generic flash-attention issue are not the same bug." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs cleanup and a version-check bug in flex_attention are unrelated." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Generic tokenizer-loading regression vs Qwen3 offset_mapping correctness; related area but different bug." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Optimizer-state corruption under FSDP/PEFT is unrelated to missing generation cache in chunked prefill." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Broken vitpose models and Ministral flash-attention failures are different model-specific issues." + }, + { + "left": "issue:42673", + "right": "issue:42796", + "accept": false, + "reason": "VRAM leak in threaded loading vs tokenizer BPE-marker handling are unrelated." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "Invalid GenerationConfig and a stale warning about num_return_sequences are different generation bugs." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Generation cache missing and SequenceFeatureExtractor padding performance are unrelated." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Unexpected init_empty_weights argument vs PretrainedConfig type-checking breakage are different v5 regressions." + }, + { + "left": "issue:42673", + "right": "issue:46655", + "accept": false, + "reason": "VRAM leak and video padding are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:46752", + "accept": false, + "reason": "Vitpose breakage and empty-list chat template crash do not share the same code path." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "PEFT version bump and flex_attention import version check are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping None and flash_attention s_aux=None crash are different defects." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "lm_head serialization inconsistency and NaNs in classification heads are not the same issue." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "GPT2 attention scaling ignored and Gemma4 bidirectional masks are separate attention-backend bugs." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and symptoms; not clearly the same bug." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GenerationConfig validation and Qwen3.5 flash-attention generation failure are different failures." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Missing Qwen3Next projections vs Qwen3.5 MoE from_pretrained error are related only by model family, not the same concrete bug." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "A testing_utils bug and a pytest-xdist race are not the same issue." + }, + { + "left": "issue:45071", + "right": "issue:46512", + "accept": false, + "reason": "Type-checking breakage and stale config.dtype are distinct issues." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "Testing-utils bug and feature-extractor padding performance are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video shape bug vs missing-torchvision AutoProcessor crash are different video-processing failures." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "flash_attention_3 training degeneration and Qwen3.5 save_pretrained regression are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping fallback and Mistral regex patch crash are separate tokenizer-loading bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 53, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:42617", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43208", + "issue:43408", + "issue:43673", + "issue:43742", + "issue:43874", + "issue:43931", + "issue:43937", + "issue:44075", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44360", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44877", + "issue:44933", + "issue:45071", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45656", + "issue:46172", + "issue:46178", + "issue:46209", + "issue:46231", + "issue:46298", + "issue:46328", + "issue:46617", + "issue:46762", + "issue:46894" + ], + "soft_pairs": [ + "issue:43408|issue:43742", + "issue:42617|issue:45341", + "issue:43122|issue:46298", + "issue:45362|issue:46328", + "issue:38617|issue:41628", + "issue:42222|issue:46178", + "issue:44877|issue:46231", + "issue:44484|issue:45310", + "issue:43937|issue:44496", + "issue:42371|issue:44261", + "issue:39401|issue:46209", + "issue:44261|issue:46762", + "issue:42371|issue:46617", + "issue:44410|issue:45478", + "issue:44933|issue:45278", + "issue:38617|issue:44220", + "issue:44075|issue:45656", + "issue:43208|issue:46894", + "issue:43066|issue:43122", + "issue:44263|issue:44485", + "issue:44483|issue:45464", + "issue:43097|issue:44493", + "issue:43673|issue:45341", + "issue:43931|issue:44410", + "issue:42222|issue:45542", + "issue:41628|issue:45071", + "issue:45071|issue:45542", + "issue:44484|issue:45478", + "issue:43874|issue:44261", + "issue:44263|issue:46172", + "issue:44360|issue:46172" + ], + "prepared_review_unit_hash": "f26fff9d5081876e11ef5b25d881d5d0b97b6ad12eb48b684f82e7be0eeb7bb8", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 35722, + "estimated_input_tokens": 8931, + "estimated_eval_tokens": 18118 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: mostly unrelated import/model/config/tokenization regressions. None of the soft pairs looks like the same concrete bug/change, so I would not merge any of them.", + "confidence": 0.9, + "canonical_issue_reason": "If one issue must stand in for the cluster, 45278 is the broadest regression report ('many import errors after update') and best reflects the most general user-facing problem among these otherwise unrelated issues.", + "canonical_pr_reason": null, + "best_issue_reason": "45278 is the most representative umbrella issue because it captures a broad cross-version regression with multiple import failures, whereas the others are narrower model- or subsystem-specific bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "Different models and failure modes: sam3 tracker/type warning vs MobileLLM key error." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "Unrelated topics: running 3d_parallel.py vs a small testing_utils bug." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenizer output regression vs noisy logging; no shared code-path bug." + }, + { + "left": "issue:45362", + "right": "issue:46328", + "accept": false, + "reason": "Chat crash on Qwen3.5 vs feature-extractor padding performance; unrelated." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different missing names/modules and likely different fixes." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "VitPose model loading breakage vs backend mapping error for tensorflow_text; different defects." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config loading vs documentation parameter mismatch; not the same issue." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "Different problems in different code paths: shard size behavior vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GenerationConfig validation for GLM-5 vs unrecognized model config; unrelated failures." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 settings API guidance vs MLA precision/config bug; not the same underlying defect." + }, + { + "left": "issue:39401", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer offset mapping bug vs FP8 inference on B200; different subsystems entirely." + }, + { + "left": "issue:44261", + "right": "issue:46762", + "accept": false, + "reason": "Both mention numerical differences, but one is a config epsilon issue and the other is a Minimax attention reference mismatch." + }, + { + "left": "issue:42371", + "right": "issue:46617", + "accept": false, + "reason": "TF32 backend guidance vs cache FileNotFoundError for trust_remote_code; unrelated." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3-related loading errors: missing projections vs from_pretrained failure." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "Shared import-error theme is too broad; one is a specific nonexistent import, the other a wide upgrade regression." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for configuration_utils vs audio feature extraction issue; no same bug." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "SGD args ignored and optimizer step called twice are different optimizer bugs." + }, + { + "left": "issue:43208", + "right": "issue:46894", + "accept": false, + "reason": "xLSTM training blockers vs CohereASR double-shift loss bug; not the same concrete change." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both tokenizer-related, but one is decoder type mismatch and the other is tokenization behavior change; not one bug." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "torch.split return values vs GLM-5 RoPE implementation; unrelated code paths." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "General chat/completions rejection in v5.3 vs streaming inference failure on Qwen3.5; not the same defect." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed tie_embeddings flag vs unexpected position-id keys; different regressions." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing in chunked_prefill vs testing_utils bug; unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Weight-shape mismatch when loading Qwen3-VL vs missing projections in qwen3next; different model-loading problems." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Model breakage in vitpose vs TensorFlow backend initialization error; not the same underlying bug." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Different import/type issues: missing AutoImageProcessor vs PreTrainedConfig type checking regression." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Type checking regression vs undefined tf backend; unrelated." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Same broad model family is not enough; max_shard_size behavior and MoE from_pretrained failure are different bugs." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different multimodal/image-processor attribute bug vs MLA epsilon precision issue." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "Different GLM/DeepSeek implementation bugs with no shared failing code-path." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "Missing ReLU in DSA indexer vs tensor transpose bug in deepseek-v4; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 54, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:27957", + "issue:38617", + "issue:42757", + "issue:42890", + "issue:43012", + "issue:43064", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43441", + "issue:43452", + "issue:43531", + "issue:43577", + "issue:43673", + "issue:43873", + "issue:43874", + "issue:43931", + "issue:43986", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44829", + "issue:44987", + "issue:45003", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45588", + "issue:45753", + "issue:46050", + "issue:46123", + "issue:46172", + "issue:46298", + "issue:46369", + "issue:46655", + "issue:46664", + "issue:46710", + "issue:46752", + "issue:46821", + "issue:46897" + ], + "soft_pairs": [ + "issue:27957|issue:43352", + "issue:43931|issue:44038", + "issue:43441|issue:44829", + "issue:43012|issue:43577", + "issue:43097|issue:43452", + "issue:45003|issue:45341", + "issue:43986|issue:45341", + "issue:44261|issue:44485", + "issue:45753|issue:46369", + "issue:44117|issue:44987", + "issue:43531|issue:45341", + "issue:43874|issue:44263", + "issue:45440|issue:46172", + "issue:43673|issue:46123", + "issue:43097|issue:45588", + "issue:44410|issue:45375", + "issue:43986|issue:45372", + "issue:44038|issue:44410", + "issue:43097|issue:46050", + "issue:38617|issue:42757", + "issue:43208|issue:46897", + "issue:43064|issue:44315", + "issue:43986|issue:44484", + "issue:46655|issue:46664", + "issue:46298|issue:46752", + "issue:46172|issue:46710", + "issue:42890|issue:46298", + "issue:44062|issue:44117", + "issue:43097|issue:43873", + "issue:38617|issue:44351", + "issue:43012|issue:46821" + ], + "prepared_review_unit_hash": "ac7cfc352ec43947854af3ca5c1fd1c6d109daf2797ad95f9e2be6e1a8672889", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36860, + "estimated_input_tokens": 9215, + "estimated_eval_tokens": 18686 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous; the soft edges are mostly false positives across unrelated model families and bug classes. No PRs are present.", + "confidence": 0.18, + "canonical_issue_reason": "issue:43097 is the broadest, most representative release-regression report in the set and has inbound references; the rest are mostly narrow model-specific failures.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the clearest general user-facing regression and the least specialized of the issues here, making it the best representative anchor for this noisy cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2, but they are different model families and likely different support paths." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Qwen3-VL checkpoint shape mismatch is not the same bug as the broader Qwen3-VL-Moe transformers v5 failure." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "One is an unsupported FlashAttention path; the other is degenerate training under flash_attention_3. Different concrete failures." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "A PyTorch precision warning on compile is unrelated to Blip2 dtype propagation staying float32." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Removing a public API is unrelated to gguf_file breaking tokenizer/model loading." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "sys.modules access in modeling_utils and a testing_utils bug are different code paths and symptoms." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "Video processor crash without torchvision is unrelated to a testing utility issue." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "LayerNorm epsilon precision errors and RoPE implementation issues are distinct model internals." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "An infinite loop on a specific Qwen3_5 image is not the same as a sam3-litetext memory overflow." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping None in from_pretrained is a different loading bug than failing on physical-intelligence/fast." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "A Qwen3-MoE sliding_window bug is unrelated to testing_utils." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Missing image-patch count method and torch.split return handling are different GLM bugs." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "DeepseekV3MoE divergence from remote impl is a different issue than the deepseek-v4 transpose/dtype bug." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Generation cache missing in chunked_prefill is unrelated to the mask pipeline's final partial batch drop." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "A tie_embeddings API removal and a flash_attention.py s_aux=None crash are different problems." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Missing Qwen3next projections and missing deepstack_visual_indexes are different model/config defects." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Missing torchvision processor crash is unrelated to the mistral_common ReasoningEffort import error." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "These affect different Qwen3 model variants with different missing pieces and failure modes." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and different dependency breakages." + }, + { + "left": "issue:43208", + "right": "issue:46897", + "accept": false, + "reason": "xLSTM training blockers and Florence2 training-loss double-shift are not the same underlying bug." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to Liger Kernel not being applied in model_init." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "A processor crash without torchvision is unrelated to the save_pretrained shard-size question." + }, + { + "left": "issue:46655", + "right": "issue:46664", + "accept": false, + "reason": "Both are SmolVLM video-processor issues, but one is padding calculation and the other is num_frames semantics." + }, + { + "left": "issue:46298", + "right": "issue:46752", + "accept": false, + "reason": "A noisy logging message and an IndexError on empty chat-template input are not the same bug." + }, + { + "left": "issue:46172", + "right": "issue:46710", + "accept": false, + "reason": "The deepseek-v4 tensor-transpose issue is not the same as the later accuracy regression after upgrade." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "Missing set_seed in integration tests is unrelated to a noisy logging message." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "AddedToken keyword duplication and TOKENIZER_MAPPING_NAMES returning None are different tokenizer-loading failures." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "A removed tie_embeddings API and offloading failures under quantization are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and likely different compatibility breaks." + }, + { + "left": "issue:43012", + "right": "issue:46821", + "accept": false, + "reason": "A bfloat16 compile warning is unrelated to TimesFM window_size raising AttributeError." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 55, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:36683", + "issue:42175", + "issue:42490", + "issue:42757", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43824", + "issue:43825", + "issue:43906", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44016", + "issue:44077", + "issue:44117", + "issue:44162", + "issue:44222", + "issue:44368", + "issue:44380", + "issue:44483", + "issue:44484", + "issue:44514", + "issue:44655", + "issue:44704", + "issue:44936", + "issue:44938", + "issue:45103", + "issue:45478", + "issue:45479", + "issue:45584", + "issue:45712", + "issue:45753", + "issue:46132", + "issue:46172", + "issue:46178", + "issue:46298", + "issue:46365", + "issue:46421", + "issue:46701" + ], + "soft_pairs": [ + "issue:44117|issue:46701", + "issue:43502|issue:43992", + "issue:44936|issue:46365", + "issue:44704|issue:45103", + "issue:44380|issue:46421", + "issue:44938|issue:46178", + "issue:43986|issue:44016", + "issue:42907|issue:45753", + "issue:42907|issue:44655", + "issue:43825|issue:44655", + "issue:43012|issue:43550", + "issue:45712|issue:46298", + "issue:43012|issue:44368", + "issue:43208|issue:43550", + "issue:43992|issue:44484", + "issue:43352|issue:43976", + "issue:43066|issue:43906", + "issue:43352|issue:44380", + "issue:44162|issue:46172", + "issue:43097|issue:46298", + "issue:45584|issue:46365", + "issue:44380|issue:46132", + "issue:44077|issue:46365", + "issue:44222|issue:45478", + "issue:36683|issue:43582", + "issue:44380|issue:45479", + "issue:42175|issue:46365", + "issue:42757|issue:43824", + "issue:43064|issue:44380", + "issue:42490|issue:43012", + "issue:44483|issue:44514", + "issue:43208|issue:43526" + ], + "prepared_review_unit_hash": "9109958f0ecc265f3b47a4c02d0f0d25c87b974d81d5f460907284dfdd3514b4", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37164, + "estimated_input_tokens": 9291, + "estimated_eval_tokens": 18838 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a duplicate cluster: the soft-similar pairs are mostly superficial keyword matches, but they point to different subsystems, APIs, or failure modes, so none should be merged.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue stands out; the items are unrelated bug reports rather than one underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a good global representative because the set spans many distinct code paths (loading, training, saving, tokenization, packaging, and model-specific failures).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44117", + "right": "issue:46701", + "accept": false, + "reason": "Tokenizer mapping fallback/import handling is unrelated to the DiffusionGemma image-text pipeline failure." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Local-files-only network leakage and a missing embed_tokens weight during from_pretrained are different loading bugs." + }, + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "train/evaluate lifecycle failure and importing Trainer being blocked by datasets are separate problems." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "cached_file kwargs forwarding and docstring processing with future annotations touch different code paths." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "GPT2 attention scaling under SDPA/FlashAttention is unrelated to chunked-prefill recompilation from cache lazy init." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python 3.14 import/load failure and tensorflow_text backend validation are not the same defect." + }, + { + "left": "issue:43986", + "right": "issue:44016", + "accept": false, + "reason": "AutoProcessor without torchvision and a notebook syntax error are completely unrelated." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals and a Qwen3.5 image loop are different model-specific issues." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Saving dequantized weights and saving Pipeline objects are both save-related but not the same underlying bug." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Bad pipeline() task messaging and save_pretrained for Pipeline objects are unrelated API issues." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A bfloat16 compilation warning is not the same as a Bamba torch.compile SDPA failure." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Leftover dummy classes and a noisy logging message are unrelated maintenance issues." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Compilation precision warnings and tie_word_embeddings warning spam come from different model behaviors." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM small-model training bugs do not match the Bamba SDPA torch.compile failure." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing checkpoint weights and the 50GB max_shard_size default concern different save/load behavior." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Flash Attention 2.0 unsupported for a model and a Python version compatibility problem are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Tokenizer decoder type handling and an isolated reproduction of another issue are not the same bug." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Unsupported FA2 on a model and GPT2 attention scaling being ignored under backends are different backend/model issues." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage and a deepseek transpose/dtype bug are separate model-specific failures." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "Removal of a public config flag and a noisy logging message are unrelated." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper empty-transcription generation failure and importing Trainer affected by datasets are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "GPT2 SDPA/FlashAttention scaling and AttentionInterface.register behavior changes are different attention-related issues." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "patchtsmixer post_init policy and Trainer import breakage have no code-path overlap." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "FP8 save_pretrained for MoE and Qwen3.5 MoE from_pretrained errors are related by model family but not the same concrete failure." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config missing vocab_size and an Apple Silicon TypeError in caching_allocator_warmup are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention scaling under SDPA/FlashAttention is not the same as a degenerate zero-loss classification bug." + }, + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "TensorFlow missing from a torch extra install and Trainer import failing because of datasets are unrelated packaging/import issues." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Offline-mode import failure from huggingface_hub and Qwen2_5_VL symbol import failure are separate missing-symbol problems." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer state corruption and GPT2 attention scaling under backend dispatch are different training vs inference bugs." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local path trailing slash warning and bfloat16 compilation precision warning are unrelated loading vs compile diagnostics." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat/completions request rejection and batched Qwen2_5_VLProcessor template crashes are different request paths." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training bugs and BeitImageProcessorFast reduce_labels truncation are unrelated model vs image-processor issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 56, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:29870", + "issue:36010", + "issue:36296", + "issue:36683", + "issue:38617", + "issue:42757", + "issue:42994", + "issue:43010", + "issue:43064", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43611", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43867", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44077", + "issue:44230", + "issue:44355", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44492", + "issue:44625", + "issue:44655", + "issue:44908", + "issue:44938", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45310", + "issue:46050", + "issue:46077", + "issue:46082", + "issue:46365", + "issue:46369", + "issue:46829", + "issue:46851", + "issue:46877" + ], + "soft_pairs": [ + "issue:43618|issue:44380", + "issue:38617|issue:43502", + "issue:43756|issue:44485", + "issue:43618|issue:43761", + "issue:42994|issue:44655", + "issue:44038|issue:44230", + "issue:43299|issue:43867", + "issue:36010|issue:42757", + "issue:44625|issue:46365", + "issue:45276|issue:46077", + "issue:43611|issue:44655", + "issue:44483|issue:46369", + "issue:43352|issue:43937", + "issue:36683|issue:43475", + "issue:44355|issue:46877", + "issue:44655|issue:46050", + "issue:36683|issue:44077", + "issue:43976|issue:44938", + "issue:44625|issue:46077", + "issue:43867|issue:45216", + "issue:46077|issue:46082", + "issue:43618|issue:45216", + "issue:36683|issue:43525", + "issue:43208|issue:46829", + "issue:36683|issue:44625", + "issue:29870|issue:36296", + "issue:43010|issue:45278", + "issue:43064|issue:44492", + "issue:43992|issue:45310", + "issue:44075|issue:46365", + "issue:44908|issue:46851" + ], + "prepared_review_unit_hash": "52a9f4a0c4adf7aaa807fb549d6b5622c8786456543306b09bcc448080d2a37e", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36574, + "estimated_input_tokens": 9144, + "estimated_eval_tokens": 18544 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a grab bag of unrelated issues. The soft pairs mostly share broad keywords like import error, config, save/load, or model family, but they do not describe the same underlying bug or fix.", + "confidence": 0.88, + "canonical_issue_reason": "No single issue is a good canonical representative here: the items span many distinct model families and failure modes, so treating any one as the duplicate target would be misleading.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue candidate stands out because none of the soft pairs are true duplicates; the safest triage is to keep them separate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Both are attention-related regressions, but one is CLIP attentions missing and the other is GPT2 attention scaling ignored under SDPA/FlashAttention; different code paths." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "An import error for configuration_utils is unrelated to local_files_only still triggering network requests." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping and GLM-5 RoPE implementation discussion are different model-specific problems." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both involve CLIP, but one is missing attentions and the other is hidden_states not returned; related area, not the same bug." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but quantized model save failure and Pipeline save_pretrained are different serialization paths." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Qwen3-VL-Moe loading regression and fp8 support for Qwen3-VL/Qwen3.5 MoE are different changes and failure modes." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3-VL-Moe loading breakage and sorted state_dict load errors are separate checkpoint-loading issues." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but GenerationMixin and is_offline_mode are unrelated symbols from different modules." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and failing to import Trainer because of datasets are unrelated." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Gemma4 resize_token_embeddings behavior and bidirectional-attention mask construction are different Gemma4 bugs." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Broken loading with base_model_prefix and Pipeline save_pretrained are separate model-loading vs pipeline-serialization issues." + }, + { + "left": "issue:44483", + "right": "issue:46369", + "accept": false, + "reason": "A chat-completions API request validation issue is unrelated to SAM3 model memory overflow." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "NemotronH FlashAttention2 support and GLM-5 GenerationConfig validation are different model/config problems." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Gemma3Config missing vocab_size and SAM3VisionEncoderOutput missing fpn_position_embeddings are unrelated attribute bugs." + }, + { + "left": "issue:44355", + "right": "issue:46877", + "accept": false, + "reason": "Compiled Python file errors and auto_docstring emitting doc-lint output at import time are not the same issue." + }, + { + "left": "issue:44655", + "right": "issue:46050", + "accept": false, + "reason": "Pipeline save_pretrained and quantized sinq round-trip breakage both touch persistence, but they are distinct code paths and bugs." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Missing vocab_size on Gemma3Config and patchtsmixer post_init validation are different configuration issues." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "These are both version-compatibility reports, but Python 3.9+ vs Python 3.14 failures are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "num_labels propagation and Gemma4 bidirectional attention masking are different configuration behaviors." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "State_dict ordering causing load errors and Qwen3.5 save_pretrained corruption are different checkpoint bugs." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Gemma4 attention-mask behavior and LlamaConfig head_dim validation are unrelated config issues." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attention output regression and Qwen3.5 saved-checkpoint regression are unrelated models and symptoms." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Gemma3Config missing vocab_size and Llama4Config missing pad_token_id are separate model config attribute problems." + }, + { + "left": "issue:43208", + "right": "issue:46829", + "accept": false, + "reason": "xLSTM training bugs and VideoPrism hidden_states mapping are unrelated model-specific issues." + }, + { + "left": "issue:36683", + "right": "issue:44625", + "accept": false, + "reason": "Both are config attribute propagation bugs, but Gemma3 vocab_size and Qwen3.5 num_labels are distinct fields in different models." + }, + { + "left": "issue:29870", + "right": "issue:36296", + "accept": false, + "reason": "BART generate() decoder_attention_mask UnboundLocalError and tensor-parallel training bug are unrelated failures." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Cache/update no_grad decoration and broad import errors after upgrade do not describe the same bug." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption and a typo in cache strategies are unrelated." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "UMT5Encoder missing embed_tokens.weight and Qwen3.5 MoE from_pretrained error are different model-loading bugs." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Ignored SGD args and inability to import Trainer due to datasets are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:46851", + "accept": false, + "reason": "Scheduler kwargs being ignored and Qwen3.5 variable-length position_ids are separate issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 57, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28282", + "issue:29870", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42890", + "issue:42898", + "issue:43010", + "issue:43064", + "issue:43502", + "issue:43575", + "issue:43576", + "issue:43618", + "issue:43723", + "issue:43824", + "issue:43867", + "issue:43874", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44246", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44734", + "issue:44855", + "issue:44938", + "issue:44964", + "issue:45071", + "issue:45125", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45600", + "issue:45850", + "issue:46132", + "issue:46202", + "issue:46285", + "issue:46617", + "issue:46664" + ], + "soft_pairs": [ + "issue:29870|issue:33453", + "issue:44938|issue:45071", + "issue:44355|issue:44855", + "issue:43867|issue:46202", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:43867|issue:45375", + "issue:43010|issue:45850", + "issue:41669|issue:44246", + "issue:36296|issue:43575", + "issue:43010|issue:43976", + "issue:43618|issue:44077", + "issue:43867|issue:45357", + "issue:43992|issue:45478", + "issue:28282|issue:33453", + "issue:43618|issue:46285", + "issue:44360|issue:45440", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:42898|issue:46664", + "issue:43502|issue:44062", + "issue:44734|issue:46617", + "issue:43723|issue:44556", + "issue:43824|issue:46617", + "issue:44222|issue:45125", + "issue:44387|issue:44556", + "issue:43502|issue:44485", + "issue:43874|issue:46617", + "issue:43576|issue:46132" + ], + "prepared_review_unit_hash": "8ec33564fd2af8b73bebc22b0f4e8f23004c44744464027d120f21018279995b", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36440, + "estimated_input_tokens": 9110, + "estimated_eval_tokens": 18476 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous issue set; the soft pairs share broad themes (loading regressions, import-time failures, tensor-parallel bugs, CI flakiness) but do not look like the same concrete underlying bug/change.", + "confidence": 0.79, + "canonical_issue_reason": "issue:33453 is the most generic, cluster-like representative here: a broad tokenizer-loading regression title that could plausibly sit at the center of the loading-related subset.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:33453 is the best overall issue representative because it is broad and concise, making it a reasonable anchor for the few loading-related items in this otherwise mixed cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29870", + "right": "issue:33453", + "accept": false, + "reason": "Different failures: BART generate() decoder_attention_mask bug vs tokenizer loading regression." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Unrelated problems: Python 3.14 import/load failure vs PretrainedConfig type-checking regression." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Different import-time issues with different triggers and code paths." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "State_dict sorting load error is not the same as Gemma pipeline roundtrip token issue." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config missing vocab_size vs FSDP2/PEFT optimizer-state corruption are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Both are flaky-test themed, but one is missing set_seed and the other is an xdist race on captured_info.txt." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Different models and different missing fields/serialization bugs." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "no_grad decoration fix vs vague broad breakage report are not the same bug." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Both mention import time, but one is import-star cleanup and the other is intermittent slowdown; too broad to merge." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug and tp OOM on loading a specific model are different concrete failures." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Different issue types: no_grad update methods vs Python/version compatibility regression." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "Different model API regressions; no shared code path is evident." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Load-state_dict error and incorrect save_pretrained visual keys are different serialization paths." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "UMT5Encoder missing embed_tokens.weight vs Qwen3.5 MoE from_pretrained error are different model-specific load failures." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "Missing PyTorch import dependency is unrelated to tokenizer-loading regression." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "CLIP attentions assignment vs CLIP text-model structure mismatch are separate regressions." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "DSA indexer activation bug and DeepseekV3MoE implementation divergence are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Test seeding flakiness vs Phi-4-multimodal model load failure are not the same bug." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Different CI/test failures with different causes." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Different CI regressions: seedless flaky integration test vs stale auto_mappings reference." + }, + { + "left": "issue:42898", + "right": "issue:46664", + "accept": false, + "reason": "Tokenization-space cleanup behavior and video num_frames semantics are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Offline local_files_only API leakage is unrelated to AddedToken keyword argument collision." + }, + { + "left": "issue:44734", + "right": "issue:46617", + "accept": false, + "reason": "KV-cache continuation indexing crash vs cache file not found are different code paths." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer loading regression and model checkpoint reload incompatibility are related only at a very high level, not the same bug." + }, + { + "left": "issue:43824", + "right": "issue:46617", + "accept": false, + "reason": "Missing import symbol and trust_remote_code cache FileNotFoundError are different issues." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained bug and missing tensor-parallel plan are unrelated model features." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization memory regression and v4-to-v5 reload failure are different failures." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only network leakage has nothing to do with GLM-5 RoPE implementation." + }, + { + "left": "issue:43874", + "right": "issue:46617", + "accept": false, + "reason": "Missing image-patch helper and trust_remote_code cache lookup failure are different multimodal loading issues." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "transformers env CLI breakage and AttentionInterface.register behavior change are separate features." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 58, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:29870", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:43010", + "issue:43065", + "issue:43232", + "issue:43299", + "issue:43475", + "issue:43502", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "issue:43901", + "issue:44222", + "issue:44230", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44749", + "issue:44908", + "issue:44938", + "issue:44961", + "issue:45071", + "issue:45125", + "issue:45362", + "issue:45375", + "issue:45468", + "issue:45478", + "issue:45850", + "issue:46143", + "issue:46144", + "issue:46209", + "issue:46617", + "issue:46664", + "issue:46726" + ], + "soft_pairs": [ + "issue:43010|issue:43065", + "issue:30333|issue:33453", + "issue:45125|issue:45478", + "issue:44355|issue:44749", + "issue:45125|issue:45375", + "issue:30990|issue:31795", + "issue:43010|issue:43232", + "issue:46664|issue:46726", + "issue:43502|issue:44961", + "issue:43502|issue:44222", + "issue:42548|issue:44351", + "issue:43827|issue:44961", + "issue:30333|issue:36296", + "issue:43299|issue:44556", + "issue:29870|issue:31795", + "issue:44355|issue:45071", + "issue:33453|issue:34689", + "issue:43576|issue:45071", + "issue:42757|issue:43576", + "issue:44315|issue:44360", + "issue:31795|issue:38617", + "issue:43576|issue:43824", + "issue:44961|issue:45468", + "issue:44485|issue:44961", + "issue:44351|issue:46617", + "issue:36296|issue:43901", + "issue:44908|issue:44961", + "issue:43576|issue:44938", + "issue:43475|issue:46617", + "issue:46143|issue:46144", + "issue:44315|issue:46209", + "issue:44222|issue:44230", + "issue:41628|issue:42548", + "issue:44485|issue:46664", + "issue:43673|issue:44492", + "issue:45362|issue:45850" + ], + "prepared_review_unit_hash": "39db6bd8a87d7dd74f1393e6fb6b24c2785293f1d2401798f0b098dd3d1e39b1", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 37785, + "estimated_input_tokens": 9447, + "estimated_eval_tokens": 19150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous cluster of unrelated Transformers issues. Only a couple of soft links look plausibly duplicate: Qwen3.5-Moe loading regressions and a pair of RoFormer attention/kwargs issues. Most other pairs are clearly different bugs, commands, docs, or model families.", + "confidence": 0.41, + "canonical_issue_reason": "issue:45125 is the clearest concrete bug report in the one strong duplicate pair here: it names the likely root cause (`_tp_plan` missing) behind the Qwen3.5-Moe loading failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45125 is the best single anchor among the items because it is specific, reproducible, and the strongest candidate for grouping with a closely related loading error report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Different subsystems: generation cache update/no-grad vs Sam3PixelDecoder dummy Conv2d." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLflow job-reporting failure and tokenizer-loading regression are unrelated." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": true, + "reason": "Both describe the same Qwen3.5-Moe loading path failing; 45125 gives the likely root cause and 45478 is a broader symptom report that could be fixed together." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled-file execution errors and a data-filtering slowdown are unrelated." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "One is tensor-parallel model-loading support; the other is a missing config field in Qwen3.5-Moe-Vision." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Sentence-Transformers loading hang and forward() documentation confusion are not the same bug." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Both are generation-related, but one is about cache update decoration and the other about sync_gpus model-kwargs handling." + }, + { + "left": "issue:46664", + "right": "issue:46726", + "accept": false, + "reason": "Video processor frame-count semantics and missing vision-backend import behavior are different issues." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "Local-files-only network leak and an unrelated 'racoon' issue do not match." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Offline-loading bug and FP8 save_pretrained/moe behavior are different code paths." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different exported symbols with no clear shared root cause." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs still mentioning pipeline() and the unrelated 'racoon' report are not duplicates." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLflow failure reporting and tensor-parallel training bug are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are version-regression loading problems, but they involve different models and failure modes." + }, + { + "left": "issue:29870", + "right": "issue:31795", + "accept": false, + "reason": "An UnboundLocalError in generate() is unrelated to forward() documentation confusion." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-file execution errors and PretrainedConfig type-checking regressions are different bugs." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer loading regression and Llama 3.2 vision model-loading breakage are different symptoms." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "Broken `transformers env` command and PretrainedConfig type checking are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing `is_offline_mode` import from huggingface_hub is not the same as the v5 env-command breakage." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger-kernel application in `model_init` and a DSA indexer/ReLU discussion do not match." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Docs confusion in model.forward and an ImportError for `layer_type_validation` are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Broken env command and Qwen2.5-VL import failure are different issues." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "The placeholder 'racoon' issue and Gemma4AudioRelPositionalEncoding bug are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation and the unrelated 'racoon' issue do not overlap." + }, + { + "left": "issue:44351", + "right": "issue:46617", + "accept": false, + "reason": "Top-level import failure for HybridCache and a cache FileNotFoundError are distinct failure modes." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training and pipeline doc wording are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs handling and the unrelated 'racoon' issue are not duplicates." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "Broken env command and Python 3.14 loading failure are different regressions." + }, + { + "left": "issue:43475", + "right": "issue:46617", + "accept": false, + "reason": "SAM3 vision encoder output attribute error and trust_remote_code cache FileNotFoundError are different problems." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": true, + "reason": "Both RoFormer issues point to the same underlying attention-path problem: methods bypassing the expected attention interface and not forwarding kwargs." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Liger-kernel model_init bug and FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained/moe behavior and FP8 inference support for Qwen3-VL/Qwen3.5-Moe are different issues." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "AutoImageProcessor import failure and PreTrainedModel import failure are separate export problems." + }, + { + "left": "issue:44485", + "right": "issue:46664", + "accept": false, + "reason": "GLM-5 RoPE implementation and SmolVLM num_frames semantics are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Generation cache missing during chunked_prefill and a typo in cache strategies are not the same bug." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A specific Qwen3.5-35B chat crash and a broad 'too many models break since 5.0' complaint are not the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The summary stays conservative about a largely heterogeneous set, and the two accepted soft links are plausible as same-bug/mergeable pairs from the titles alone. No obvious overstatement beyond the evidence in the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 59, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43673", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44230", + "issue:44246", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45278", + "issue:45341", + "issue:45468", + "issue:45478", + "issue:45542", + "issue:45753", + "issue:45850", + "issue:46095", + "issue:46218", + "issue:46328", + "issue:46655", + "issue:46664" + ], + "soft_pairs": [ + "issue:44273|issue:45542", + "issue:44623|issue:45230", + "issue:44360|issue:46664", + "issue:44273|issue:45003", + "issue:44038|issue:45125", + "issue:44393|issue:45071", + "issue:44393|issue:45753", + "issue:41669|issue:43673", + "issue:43576|issue:44556", + "issue:43525|issue:45850", + "issue:43572|issue:45850", + "issue:44556|issue:45092", + "issue:44492|issue:45850", + "issue:45278|issue:45850", + "issue:42371|issue:44492", + "issue:43408|issue:44315", + "issue:41669|issue:46218", + "issue:44230|issue:46095", + "issue:44246|issue:44273", + "issue:42890|issue:44908", + "issue:43576|issue:44908", + "issue:43352|issue:44393", + "issue:43976|issue:44393", + "issue:44908|issue:45071", + "issue:44295|issue:45230", + "issue:44492|issue:46328", + "issue:44297|issue:45230", + "issue:44704|issue:44908", + "issue:41669|issue:44351", + "issue:41669|issue:45341", + "issue:43976|issue:45850", + "issue:43976|issue:46218", + "issue:44393|issue:45478", + "issue:44077|issue:44861", + "issue:31795|issue:33453", + "issue:43576|issue:44273", + "issue:44315|issue:44829", + "issue:43519|issue:46655", + "issue:44360|issue:45468", + "issue:44273|issue:44908" + ], + "prepared_review_unit_hash": "45b944b6bf04f2bff0f0bcfae943d23cb82e55976c5d0e73f3eeeb884996b73b", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 40, + "serialized_chars": 39102, + "estimated_input_tokens": 9776, + "estimated_eval_tokens": 19808 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is noisy and mostly contains unrelated regressions. The strongest common thread is import/lazy-loading-related issues, but the provided pairs do not look like exact duplicates of the same concrete bug/change.", + "confidence": 0.46, + "canonical_issue_reason": "issue:44273 is the most representative broad bug in the set, centered on lazy loading/import behavior, which is the closest shared theme across the few related entries.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44273 is the best global issue candidate because it names the underlying subsystem most clearly and could plausibly anchor nearby import/lazy-loading regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Both are import/runtime-related, but one is a lazy-loading problem while the other is a TF backend/tensorboard-specific error; not the same bug." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "One is a specific processor.save_pretrained missing-files bug; the other is an uninformative generic bug report." + }, + { + "left": "issue:44360", + "right": "issue:46664", + "accept": false, + "reason": "Completely different domains: DSA indexer/ReLU vs SmolVLM video padding/num_frames." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Both touch import/lazy-loading behavior, but the concrete failures differ and there is not enough evidence of the same underlying defect." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Same model family, but one is a general Qwen3-VL-Moe bug and the other is a tensor-parallel _tp_plan issue." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL output bug vs a PretrainedConfig type-checking regression; unrelated code paths." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Both mention Qwen3-VL/Qwen3.5 vision models, but the symptoms and likely root causes are different." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-time performance/lazy-loading is the only overlap; one is about import * usage, the other about chunked_prefill cache presence." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Both are v5 regressions, but in different surfaces: the env command vs checkpoint reload compatibility." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "A missing pad_token_id on one config is much more specific than the broad 'breaks too many models' complaint." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "A missing pad_token_idx on StableLmConfig is a specific config bug, not the same as a broad version-wide complaint." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both involve v5 upgrade/loading problems, but they affect different model families and likely different code paths." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A cache-strategy typo is unrelated to a generic version regression complaint." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Broad import errors after upgrade are not the same concrete issue as a general 'too many models' complaint." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API settings and cache-strategy typo are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Model-type warning vs Liger Kernel application bug are different problems." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Both are import-related, but one is about import * and performance; the other is a missing module import for a specific class." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support for Qwen3-VL/Moe is unrelated to a DeepSeek weight-conversion question." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Both concern import/lazy-loading performance, but one is slow import timing and the other is a functional lazy-loading failure; not clearly the same bug." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test seed flakiness and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Different subsystems entirely: env command vs scheduler kwargs." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Nemotron Flash Attention support and Qwen3-VL bounding-box output are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python-version compatibility and Qwen3-VL output issues are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler keyword handling is unrelated to PretrainedConfig type checking." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Specific position_ids buffer access bug vs generic bug report." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Both are efficiency/cleanup issues, but one is a typo and the other is a padding conversion performance regression." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer_class mismatch in save_pretrained is specific; the other item is generic and non-specific." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "Processor kwargs forwarding and scheduler kwargs forwarding are different APIs and bugs." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Both involve imports, but one is performance/lazy-loading and the other is a missing symbol export/import error." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import-time performance issue and a testing_utils bug are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Python version support regression is not the same as a broad 'many models break' complaint." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python compatibility and a specific module import failure are unrelated." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3-family issues with different symptoms and model paths; not the same concrete bug." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Optional post_init validation and tied-weights key handling are separate implementation issues." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion in model.forward is unrelated to tokenizer loading regression." + }, + { + "left": "issue:41669", + "right": "issue:44273", + "accept": false, + "reason": "Import * overhead and lazy-loading failure may be loosely related, but they are not clearly the same underlying change or bug." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel not applied with model_init and flash_attention_3 training degeneration are different issues." + }, + { + "left": "issue:43519", + "right": "issue:46655", + "accept": false, + "reason": "Different video-processor bugs in different models: timestamp calculation vs padding calculation." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer/ReLU discussion is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy-loading regression and inverse_sqrt scheduler kwargs are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 60, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:28348", + "issue:29870", + "issue:30333", + "issue:31515", + "issue:31795", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41669", + "issue:42371", + "issue:42913", + "issue:42936", + "issue:43066", + "issue:43298", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43606", + "issue:43618", + "issue:43704", + "issue:43717", + "issue:44230", + "issue:44263", + "issue:44351", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44655", + "issue:44861", + "issue:44908", + "issue:44938", + "issue:44998", + "issue:45230", + "issue:45250", + "issue:45290", + "issue:45310", + "issue:45468", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46144", + "issue:46172", + "issue:46218", + "issue:46894" + ], + "soft_pairs": [ + "issue:45468|issue:46894", + "issue:45230|issue:45310", + "issue:43576|issue:46218", + "issue:43519|issue:44485", + "issue:44655|issue:44861", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689", + "issue:43704|issue:44485", + "issue:46018|issue:46172", + "issue:44485|issue:45468", + "issue:42371|issue:43704", + "issue:43502|issue:43519", + "issue:45850|issue:46218", + "issue:43519|issue:44623", + "issue:44263|issue:45468", + "issue:43606|issue:45468", + "issue:43519|issue:43704", + "issue:43298|issue:46095", + "issue:43352|issue:46218", + "issue:44351|issue:46218", + "issue:42913|issue:43066", + "issue:28348|issue:45250", + "issue:44908|issue:44998", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44998|issue:45850", + "issue:44938|issue:44998", + "issue:44861|issue:45290", + "issue:31515|issue:38617", + "issue:43572|issue:44861", + "issue:43525|issue:44861", + "issue:44230|issue:44623", + "issue:28348|issue:43352", + "issue:29870|issue:31515", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:31515|issue:31795", + "issue:44230|issue:45310", + "issue:43298|issue:44230", + "issue:43717|issue:46144", + "issue:42936|issue:46018" + ], + "prepared_review_unit_hash": "3265befcec5f5e61b44bd85dc3dd95a244ecc1c861db1d2924791abc1df50407", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 42, + "serialized_chars": 38806, + "estimated_input_tokens": 9702, + "estimated_eval_tokens": 19660 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is broadly heterogeneous, with a loose v5/regression theme but no true duplicate pairs among the soft edges. No PRs are present. If one issue must serve as the anchor, 43066 is the most concrete and representative of the tokenizer-v5 subset.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 43066 is the clearest concrete bug report here: a specific v5 tokenizer decoder-type regression.", + "canonical_pr_reason": null, + "best_issue_reason": "43066 is the most suitable representative because it is specific, actionable, and on the main recurring v5/tokenizer theme; most other issues are either vague placeholders or unrelated subsystems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45468", + "right": "issue:46894", + "accept": false, + "reason": "Both are audio-related, but they describe different model bugs and different code paths." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "Both are generic loading failures, but one is a placeholder report and the other is a Qwen3.5 MoE-specific from_pretrained error." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Both involve v5 breakage, but one is the `transformers env` command and the other is a missing model import symbol." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Processor timestamp calculation and GLM-5 RoPE implementation are different bugs in different components." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Both touch save/pretrained internals, but Pipeline saving and tied-weight-key handling are distinct failures." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp calculation in a processor and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star cleanup/performance work is unrelated to a VRAM leak in Qwen3ForCausalLM." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion about inputs/past_key_values is not the same as a model-loading regression." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "These are different failures in different model implementations: memory leak versus RoPE math." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeek-V4 shared-expert gating and DeepSeek-V4 tensor transpose issues are distinct bugs." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Different models and different implementation areas; not the same underlying defect." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 API settings and a VRAM leak are unrelated issues." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Offline-loading network requests and processor timestamp math are unrelated." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "A broad complaint that v5 breaks many models is not the same bug as a specific Qwen3_5 import failure." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Processor timestamp logic and missing save_pretrained files are different code paths." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "GLM MoE indexing and Gemma-4 audio positional encoding are unrelated bugs." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU-offload device mismatch and audio positional encoding are different problems." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Processor timestamp calculation and VRAM leak symptoms are unrelated." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Continuous batching/audio support and DeepSeek-V4 weight conversion are not the same change." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention support for NemotronH and a Qwen3_5 import error are different issues." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and different modules." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both are tokenizer v5 regressions, but one is a general behavior change and the other is a specific decoder-type mismatch; not clearly one bug." + }, + { + "left": "issue:28348", + "right": "issue:45250", + "accept": false, + "reason": "They only share the Flash Attention 2.0 wording; one is GPT2 support and the other is a generic title." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs handling is unrelated to the placeholder issue title." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A typo in cache strategies is unrelated to the placeholder issue." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attention assignment regression is unrelated to the placeholder issue." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "A placeholder title does not indicate the same underlying bug as the broad v5-breakage complaint." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 loading failure is unrelated to the placeholder issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight key crashes and chat-template tool-call crashes are different code paths." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained loading and a missing import symbol are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad-token regression and tied-weight-key handling are different bugs." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Llama4Config pad_token_id absence is unrelated to tied-weight-key crashes." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support and missing processor save files are unrelated." + }, + { + "left": "issue:28348", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2.0, but they concern different models and different symptoms." + }, + { + "left": "issue:29870", + "right": "issue:31515", + "accept": false, + "reason": "BART generate mask error and slow checkpoint loading are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow reporting failure and slow loading are unrelated issues." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint loading performance and tensor-parallel training are different problems." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Loading performance and documentation confusion do not describe the same bug." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 support request and Qwen3.5 MoE loading error are different concerns." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models and FP8 inference support are unrelated." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init-weights behavior and RoFormer attention interface usage are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2former ignore_value handling and DeepSeek-V4 shared-expert gating are distinct bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 24, + "review_unit_index": 61, + "review_unit_count": 61, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-502", + "nodes": [ + "issue:37934", + "issue:42936", + "issue:43240", + "issue:43668", + "issue:43813", + "issue:44031", + "issue:46144" + ], + "soft_pairs": [ + "issue:43668|issue:44031", + "issue:42936|issue:43240", + "issue:43240|issue:46144", + "issue:42936|issue:44031", + "issue:42936|issue:46144", + "issue:42936|issue:43813", + "issue:43240|issue:43668", + "issue:43668|issue:43813", + "issue:37934|issue:43668", + "issue:42936|issue:43668" + ], + "prepared_review_unit_hash": "01756314d74ff7fd2bafa898ce619016749025cbeb37382eaabab76eeb6ef34c", + "component_budget": { + "node_count": 502, + "item_count": 502, + "soft_pair_count": 1845, + "serialized_chars": 957264, + "estimated_input_tokens": 239316, + "estimated_eval_tokens": 478888 + }, + "budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 10, + "serialized_chars": 6932, + "estimated_input_tokens": 1733, + "estimated_eval_tokens": 3722 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set: the issues span unrelated models/components (norm behavior, Mask2Former loss handling, ModernBERT typing, tokenizer warnings, RoFormer attention plumbing, etc.). None of the soft pairs look like the same underlying bug.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No single canonical issue fits this cluster because the items are unrelated; if one must be highlighted, issue:43240 is the most concrete and broadly discussed bug report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "Different problems: a type-hint bug in ModernBERTConfig versus a tokenizer regex warning regression." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2Former ignore_value handling is unrelated to kwargs not being forwarded into a loss function." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss kwarg forwarding and RoFormer attention interface usage are separate code paths and bug types." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former loss/ignore_value behavior does not match tokenizer regex warning behavior." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Mask2Former ignore_value logic and RoFormer attention interface wiring are unrelated model bugs." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "A missing ignore_value use is unrelated to a typo in a quantization attribute name." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Passing kwargs to loss calculation is unrelated to an incorrect type hint for norm_eps." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "ModernBERTConfig typing and a quantization-attribute typo are different issues with no shared code-path." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Llama4TextL2Norm vs RMS norm semantics is a model implementation question, not a ModernBERTConfig type-hint bug." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value handling and ModernBERTConfig norm_eps typing are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10519-3", + "nodes": [ + "issue:10519", + "issue:11005", + "issue:11307" + ], + "soft_pairs": [ + "issue:10519|issue:11307", + "issue:10519|issue:11005", + "issue:11005|issue:11307" + ], + "prepared_review_unit_hash": "1b7be70740fefbce69d0859117844a270100235e1da5d3cd02219b9791650827", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2877, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2877, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three issues are not duplicates: they concern unrelated topics (tokenizer truncation behavior, a ReduceLROnPlateau-like scheduler feature, and Wav2Vec2 word boundary offsets).", + "confidence": 0.98, + "canonical_issue_reason": "Issue #10519 is the most concrete, self-contained feature request in the set; the other two are unrelated problems and do not form a single underlying bug/feature.", + "canonical_pr_reason": null, + "best_issue_reason": "#10519 is the best representative only in the sense that it is the clearest and most specific issue, but the cluster does not appear to be a true duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different areas and requests: tokenizer truncation direction vs. Wav2Vec2 word timestamp offsets. No shared underlying bug or change." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Token truncation behavior is unrelated to ReduceLROnPlateau-like scheduling functionality." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "A learning-rate scheduler feature request and Wav2Vec2 word offset extraction are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 3, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-13244-3", + "nodes": [ + "issue:13244", + "issue:15980", + "issue:16618" + ], + "soft_pairs": [ + "issue:13244|issue:16618", + "issue:15980|issue:16618" + ], + "prepared_review_unit_hash": "38bf940c4f209755e2acdf93dabff257fb52cd5ea7a99afcc7be398aaeab97fe", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not true duplicates: one is about TAPAS tokenization, one about private-model download/auth error messaging, and one about a generic BERT loading failure. The similarities are superficial and point to different code paths.", + "confidence": 0.87, + "canonical_issue_reason": "Issue #16618 is the broadest and most central user-facing failure in the set, so it works best as the anchor even though it is not actually a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "#16618 is the most generally phrased and likely most reusable report among the three, but it still describes a different problem than the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "TAPAS tokenization mismatch is a model-specific preprocessing issue, not the same bug as a generic \"can't load bert-base-uncased\" failure." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Private-model download/auth error messaging is about access-control handling, while #16618 is a general model-loading problem for a public model." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 4, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-26350-2", + "nodes": [ + "issue:26350", + "issue:33260" + ], + "soft_pairs": [ + "issue:26350|issue:33260" + ], + "prepared_review_unit_hash": "90e87c729a81dd1f09e3b5ddc072c4bb0c5f03d5f26cee537a643ffe876d8212", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate feature requests in the same general area (community support for additional model/attention architectures), but they target different capabilities and different code paths: Flash Attention 2 vs GGUF. They should not be deduplicated.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No single canonical issue fits both, since the requests are about different features; if one must be chosen as a representative issue, issue 26350 is slightly more central as an earlier, specific architecture-support request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different underlying requests: one is for Flash Attention 2 support, the other for GGUF support. They touch different integration paths and are not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-27088-9", + "nodes": [ + "issue:27088", + "issue:28936", + "issue:38435", + "issue:43834", + "issue:44365", + "issue:44370", + "issue:46222", + "issue:46226", + "issue:46931" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:27088|issue:38435", + "issue:27088|issue:28936", + "issue:28936|issue:43834", + "issue:28936|issue:44365", + "issue:28936|issue:38435", + "issue:28936|issue:44370", + "issue:38435|issue:44370", + "issue:38435|issue:44365", + "issue:38435|issue:43834", + "issue:38435|issue:46222", + "issue:38435|issue:46226", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:43834|issue:46222", + "issue:43834|issue:46226", + "issue:43834|issue:46931", + "issue:44365|issue:44370", + "issue:44365|issue:46222", + "issue:44365|issue:46226", + "issue:44365|issue:46931", + "issue:44370|issue:46222", + "issue:44370|issue:46226", + "issue:44370|issue:46931", + "issue:46222|issue:46226", + "issue:46222|issue:46931", + "issue:46226|issue:46931" + ], + "prepared_review_unit_hash": "88e0a639e03a60463eb0cea23db487aeb708e89788e6a7f92aa8185b4849e3bc", + "component_budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 29, + "serialized_chars": 15981, + "estimated_input_tokens": 3996, + "estimated_eval_tokens": 8248 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 29, + "serialized_chars": 15981, + "estimated_input_tokens": 3996, + "estimated_eval_tokens": 8248 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated language-specific translation requests plus one generic placeholder-template subgroup and one separate Thai duplicate pair. The safest duplicate calls are the template-title issues ([i18n-] Translating docs to ) and the identical Thai-title pair; the Turkish, Spanish, and Romanian requests should stay separate because they target different languages.", + "confidence": 0.88, + "canonical_issue_reason": "43834 is the cleanest representative of the generic placeholder-template subgroup: it is the earliest of those issues and serves as the best anchor for that duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "43834 is the best overall issue representative for the main duplicate subgroup; the other high-similarity items are either the same template issue or separate language requests that should not be merged.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different language-specific translation request versus the generic placeholder template; the similarity is thematic, not the same issue." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Turkish request and generic template issue are not the same underlying task." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Turkish-specific translation request versus template placeholder; not a duplicate." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Different target languages, so these are separate translation requests." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish and Spanish translation requests are distinct." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish-specific request versus the generic template issue; not the same underlying issue." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different language request and template placeholder; similarity is too broad to merge." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish and Romanian translation requests are separate." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Different language request versus template placeholder; not a duplicate." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian request versus generic placeholder template; only the issue pattern overlaps." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Different concrete language request; the shared structure is not enough to merge." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian-specific request versus template placeholder; not the same issue." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Different language/request content; the similarity is only at a broad template level." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Different language/request content; not a duplicate." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Same generic placeholder-style title and very high similarity signals; body text is not shown, so this is a likely duplicate rather than a certain exact match." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Same generic placeholder-style title and very high similarity signals; likely duplicate based on the available metadata." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "The titles and apparent language context do not line up; the shared similarity is not enough to treat these as the same issue." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Only template-level similarity is visible; not enough evidence that these are the same request." + }, + { + "left": "issue:43834", + "right": "issue:46931", + "accept": false, + "reason": "Different title and likely different underlying request; not a duplicate." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Same placeholder-style title and extremely high similarity signals; likely duplicate, though the body text is not exposed here." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Different language/request content; template similarity alone is not enough." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Different language/request content; not clearly the same issue." + }, + { + "left": "issue:44365", + "right": "issue:46931", + "accept": false, + "reason": "Different title/content and no clear duplicate relationship." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Template similarity only; the visible metadata does not support the same underlying request." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Template similarity only; not enough evidence to merge." + }, + { + "left": "issue:44370", + "right": "issue:46931", + "accept": false, + "reason": "Different issue content; not a duplicate." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": true, + "reason": "Same exact Thai title and very high similarity signals; likely duplicate based on metadata, though the body text is not shown." + }, + { + "left": "issue:46222", + "right": "issue:46931", + "accept": false, + "reason": "Different titles and no strong evidence they describe the same request." + }, + { + "left": "issue:46226", + "right": "issue:46931", + "accept": false, + "reason": "Different titles and no strong evidence of a duplicate relationship." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The clustering is broadly plausible, but the writeup is a bit too strong for the packet. The packet shows identical or very similar titles, body lengths, and soft-similarity signals, but it does not expose the body text, so phrases like \"exact duplicate\" and \"near-identical body\" go beyond what is directly supported. The conservative parts are fine, especially separating the clearly different language requests, but the accepted duplicate verdicts should be phrased as likely duplicates rather than certain exact matches." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36718-19", + "nodes": [ + "issue:36718", + "issue:36817", + "issue:36879", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322", + "issue:46133", + "issue:46645", + "issue:46653" + ], + "soft_pairs": [ + "issue:36718|issue:41115", + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:36879|issue:37015", + "issue:37015|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41115|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42738|issue:45027", + "issue:42738|issue:46653", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:43909|issue:46645", + "issue:44121|issue:45322", + "issue:43837|issue:45027", + "issue:37015|issue:45027", + "issue:43837|issue:45177", + "issue:45322|issue:46133", + "issue:42503|issue:46653", + "issue:45027|issue:46653" + ], + "prepared_review_unit_hash": "16e60a44c56da7a4bbce2e198033acf1239d33e26d05d0d5b370a59dedf8da9e", + "component_budget": { + "node_count": 19, + "item_count": 19, + "soft_pair_count": 40, + "serialized_chars": 24672, + "estimated_input_tokens": 6168, + "estimated_eval_tokens": 12592 + }, + "budget": { + "node_count": 19, + "item_count": 19, + "soft_pair_count": 40, + "serialized_chars": 24672, + "estimated_input_tokens": 6168, + "estimated_eval_tokens": 12592 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These items are mostly unrelated model-request issues. Despite title-level similarity, they do not share the same concrete bug or code change, so no soft edge should be accepted.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36718", + "right": "issue:41115", + "accept": false, + "reason": "Different domains: keypoint matching vs a MiniCPM3 architecture request." + }, + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Unrelated changes: EuroBert config entry vs a DEIMv2 model addition." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "Different capabilities: EuroBert config vs Qwen3-TTS support." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "Different model families: EuroBert config vs RF-DETR model add." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "Both are model additions, but for different vision models and code paths." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "Detection model request vs audio model request; not the same change." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "Different model families: RF-DETR vs weight-sparse transformer support." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "Vision detection vs speech ASR; unrelated underlying change." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "Different modalities and models: RF-DETR vs Qwen3-TTS." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are unrelated model additions." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "BERT-family model request vs weight-sparse transformer request." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "Text LM request vs audio model request." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "Different model additions; no shared code path or exact target." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT vs ModernVBERT are separate model integrations." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "Architecture request vs TTS support; not the same bug/change." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 architecture vs sparse-transformer support are different changes." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "Different model families and integration points." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 architecture request vs Voxtral TTS support." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "Vision detection request vs ASR support; different pipeline." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT vs ASR are unrelated model requests." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT config/support vs Qwen3-TTS; different models and modalities." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "Generic BERT-RoPE support vs ModernVBERT model addition." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT/RoPE support vs Voxtral TTS; unrelated." + }, + { + "left": "issue:42738", + "right": "issue:46653", + "accept": false, + "reason": "Model request vs unrelated person-named issue; no common change." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are different modalities and likely separate implementations." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Two different model requests with different integration points, not one concrete change." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "TTS support vs Voxtral TTS model addition; different models and code paths." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "TTS support vs OCR model addition; unrelated." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "ASR support vs audio model request; different model integrations." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR support vs EUPE; unrelated model families." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "ASR support vs sparse-transformer request; different changes." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "Audio model vs SAM3-LiteText; not the same bug/change." + }, + { + "left": "issue:43909", + "right": "issue:46645", + "accept": false, + "reason": "Audio model vs SWA+sink Granite option; different model families." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Sparse transformer feature vs EUPE model request." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "ASR model request vs Voxtral TTS; different modalities and implementations." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT vs Voxtral TTS; unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "ASR support vs DeepSeek-OCR2; separate integrations." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are distinct model requests." + }, + { + "left": "issue:42503", + "right": "issue:46653", + "accept": false, + "reason": "Model request vs unrelated person issue." + }, + { + "left": "issue:45027", + "right": "issue:46653", + "accept": false, + "reason": "TTS model request vs unrelated person issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-38740-2", + "nodes": [ + "issue:38740", + "issue:46443" + ], + "soft_pairs": [ + "issue:38740|issue:46443" + ], + "prepared_review_unit_hash": "5cd8c890d4bcc179e9c819b1ea762486136795be4df327e465181721a0451803", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one is a documentation request for adding Pruna as an optimization framework, while the other is a Vision RoPE refactor. They should not be clustered as duplicates or the same underlying task.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38740", + "right": "issue:46443", + "accept": false, + "reason": "Different topics and likely different workstreams: docs/framework listing vs. vision model positional-encoding refactor. No evidence they refer to the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41084-2", + "nodes": [ + "issue:41084", + "issue:43037" + ], + "soft_pairs": [ + "issue:41084|issue:43037" + ], + "prepared_review_unit_hash": "2e017f9c21fcf03e45dba68ae8edfb487c99f75213210b9e0ad96076db556603", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are about different topics: one is about set block decoding, the other about DeepSeek v3.2 support. They do not appear to describe the same underlying problem or change.", + "confidence": 0.97, + "canonical_issue_reason": "No true canonical issue: these are unrelated requests. If forced to pick a representative, issue 43037 is the clearer standalone feature request, but it is not a duplicate of 41084.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43037 is the better standalone representative because it is a clearly scoped support request, but it should not be treated as the same issue as 41084.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different subjects and likely different code paths; no evidence they are the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 9, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41306-2", + "nodes": [ + "issue:41306", + "issue:41826" + ], + "soft_pairs": [ + "issue:41306|issue:41826" + ], + "prepared_review_unit_hash": "5462f526d02d5e0e3ddbd4ee8e2a8198994a2e4eea164796fb0c2c215df0d4dc", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are both Tiled-related feature requests, but they target different components and different changes: one asks for a `from_pretrained()` flag for TiledFusedLogitsLoss/LigerFusedLinearCrossEntropy, the other proposes integrating TiledMLP for lower memory usage. They do not look like duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "Issue 41306 is the more concrete, narrowly scoped API request; it is easier to treat as the representative item for this small cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "41306 is the best single issue to keep as the cluster representative because it is a specific request with a clear entry point (`from_pretrained()`), while 41826 is a separate feature proposal.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41306", + "right": "issue:41826", + "accept": false, + "reason": "Reject: they concern different features and code paths. 41306 is about a FLCE/TiledFusedLogitsLoss `from_pretrained()` flag, while 41826 is about adding TiledMLP for memory reduction; same subsystem name is not enough to make them the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43598-2", + "nodes": [ + "issue:43598", + "issue:43600" + ], + "soft_pairs": [ + "issue:43598|issue:43600" + ], + "prepared_review_unit_hash": "5b1743bdf6fa6ab995b85eda40351d80651905dc883fbb2107b689a8d35203cb", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1583, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1583, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both issues are about custom trainers in `Trainer`, but they target different behaviors: one is about `remove_unused_column` customizability, the other about FLOPS estimation. They are related in theme, not the same underlying bug.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43598 is the broader, more representative Trainer customizability problem and best captures the cluster\u2019s theme.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43598 is the best overall representative because it is the more general Trainer customization concern, while 43600 is a narrower metrics-specific case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "They share the broad custom-trainer context, but the concrete problems differ: one concerns `remove_unused_column` behavior, the other FLOPS counting. These would not naturally be fixed by one patch." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43684-2", + "nodes": [ + "issue:43684", + "issue:44405" + ], + "soft_pairs": [ + "issue:43684|issue:44405" + ], + "prepared_review_unit_hash": "d1d71c24d3ea322e6b91dd260f6631d1650e883cb35da46a39ceefe32ef3caca", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not duplicates: one requests Qwen3-Omni auto-model registration, while the other requests AutoModelForSequenceClassification support for Qwen3.5. They concern different model families and different auto-class paths.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying gaps: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 sequence-classification support. Not the same code path or mergeable as one fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43835-4", + "nodes": [ + "issue:43835", + "issue:43845", + "issue:43846", + "issue:43859" + ], + "soft_pairs": [ + "issue:43835|issue:43845", + "issue:43835|issue:43846", + "issue:43835|issue:43859", + "issue:43845|issue:43846", + "issue:43845|issue:43859", + "issue:43846|issue:43859" + ], + "prepared_review_unit_hash": "2e8fae2a130039a44e0e964c6e9ac4e358b2e3d569655d8d6fe9d95e2d0659ab", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster looks like weakly related or possibly spammy issue reports, with three nearly identical \"huggingface\" issues and one unrelated \"Fett\" issue. The evidence is too sparse to confidently collapse any pair, so I would not accept any soft-edge merges.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 43845 is the best canonical candidate because it is the earliest of the near-identical \"huggingface\" reports and matches the strongest duplicate signal, but the available metadata is too thin to be fully confident.", + "canonical_pr_reason": null, + "best_issue_reason": "43845 is the best representative issue for the near-duplicate subgroup since it is earliest and has the closest match to 43846/43859 by title and metadata. Still, the lack of body content makes this only a tentative choice.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "The titles are different and there is no body or target evidence tying them to the same underlying issue." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "The only signal is weak generic similarity; there is no concrete evidence they report the same bug." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "No meaningful evidence beyond soft similarity, and the titles do not indicate the same problem." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": false, + "reason": "They are near-identical on title and timing, but the missing descriptions make it impossible to confirm they are the same underlying issue." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "Same generic title, but no concrete body or target details to prove they are duplicates." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "The metadata is too sparse to establish that these are the same report rather than separate vague issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44371-5", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:44937", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "fd10da3960dbc3b15aff67690cd496a20ffd022cb2f47e157c3f240ecd422fb0", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are mostly low-information or spam-like issues with only superficial title similarity. No pair has enough concrete evidence to treat as the same underlying problem.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 44418 is the only one that reads like a plausible legitimate ticket ('Documentation Enhancement Suggestion'); the others are generic spam or likely spam.", + "canonical_pr_reason": null, + "best_issue_reason": "44418 is the best issue to keep as the representative because it is the most substantive and specific of the set, even though it is still vague.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are titled '', but that is not evidence of the same underlying issue; they are just similarly labeled spam placeholders." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Same as above: identical spam titles, but no concrete bug or change described." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam-like and the other is a vague documentation suggestion; no shared concrete problem." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "'' and 'Check out \"Google LLC\"' look like unrelated spam/ad content, not the same issue." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "A documentation enhancement request is not the same as a spam placeholder issue." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "No shared substance beyond both being vague; one is a doc suggestion, the other is spam." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are titled '', but there is no evidence they describe the same report rather than two separate spam submissions." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus promotional-looking title; unrelated." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus promotional-looking title; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45145-2", + "nodes": [ + "issue:45145", + "issue:45151" + ], + "soft_pairs": [ + "issue:45145|issue:45151" + ], + "prepared_review_unit_hash": "caab7c9893099df5958def6996728a95f1d7e5877631e18d9fb920580001910e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "#45145 and #45151 look like likely duplicates based on the exact same title, close body lengths, and the soft similarity signal. The metadata does not prove identical bodies, so the assessment is kept conservative. #45145 is the better canonical issue because it has an inbound reference and is slightly more established.", + "confidence": 0.86, + "canonical_issue_reason": "#45145 is the better canonical issue because it has the exact same title as #45151 and already has an inbound reference, giving it a slight edge as the anchor record. The available metadata suggests duplication, but does not let us verify identical bodies.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the best issue to keep: it matches #45151 exactly on title, has a nearby body length, and already has an inbound reference. That makes it the more suitable surviving record for deduplication.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Accept as likely duplicate: the titles match exactly and the soft similarity signal supports the same underlying report. The evidence is limited to metadata, so this is a cautious acceptance rather than a claim of identical bodies." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The conclusion is directionally plausible, but it overstates the evidence. The packet shows identical titles, similar body lengths, and a soft similarity signal, but not that the bodies are the same or that the issues are \"near-identical.\" Also, \"same title/body\" in the canonical_issue_reason is unsupported because the body text is not provided. Please keep the verdict more conservative and grounded in the available metadata." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45762-2", + "nodes": [ + "issue:45762", + "issue:45763" + ], + "soft_pairs": [ + "issue:45762|issue:45763" + ], + "prepared_review_unit_hash": "2f2b407273cff70093aed772b7642e21a24c582ad0bc47193905469652fc0c3d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues are just short name/date strings and do not describe the same bug or request. They are not duplicates or a meaningful cluster match.", + "confidence": 0.98, + "canonical_issue_reason": "Issue #45762 is marginally more complete as a standalone entry, but neither issue contains actionable problem description.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #45762 is the least ambiguous of the two, though both are essentially non-technical placeholder-style issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "The titles are different personal names with different dates and no shared issue context, code path, or explicit linkage; this does not look like the same underlying bug/request." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 17, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45764-2", + "nodes": [ + "issue:45764", + "issue:45765" + ], + "soft_pairs": [ + "issue:45764|issue:45765" + ], + "prepared_review_unit_hash": "d2eedcf3f6d0d9123becb11782b6bc8500a32e97c52b3bf2a50d07e5f89ceca4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items look like unrelated low-content issues containing personal names/dates, not duplicate bug reports or change requests. The soft similarity is likely due to shared surname only.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good canonical choice: both are empty, non-technical, and appear unrelated apart from the shared surname.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "Different people/names and dates; no shared bug, request, or code-path context, so this is not a duplicate cluster." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 18, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45810-2", + "nodes": [ + "issue:45810", + "issue:46295" + ], + "soft_pairs": [ + "issue:45810|issue:46295" + ], + "prepared_review_unit_hash": "bcaf14357401832679d17c8feee8437dfa5ff8ea6d93847ef44eca4f42f0a906", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate feature requests: one for adding Qwen3_5 token classification/value-model support, and one for FP8 training support. They do not describe the same underlying bug or change.", + "confidence": 0.97, + "canonical_issue_reason": "Not a duplicate cluster; if one must serve as the representative issue, #46295 is broader and already has more inbound attention.", + "canonical_pr_reason": null, + "best_issue_reason": "#46295 is the better standalone representative because it has a clearer, broader training-focused request and more inbound references.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "Different feature areas and code paths: model head/value-model support vs FP8 training support. They are not the same underlying request." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 19, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45869-5", + "nodes": [ + "issue:45869", + "issue:45870", + "issue:45871", + "issue:45872", + "issue:45873" + ], + "soft_pairs": [ + "issue:45869|issue:45870", + "issue:45869|issue:45871", + "issue:45870|issue:45871", + "issue:45870|issue:45872", + "issue:45870|issue:45873", + "issue:45871|issue:45872", + "issue:45871|issue:45873", + "issue:45872|issue:45873" + ], + "prepared_review_unit_hash": "2ae9b1dd413645e4c41db0b3c1423eafab2e0d5b163f327cedeba9a1926e3edc", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5728, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5728, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The set splits into two exact duplicate issue pairs: 45870\u219445871 and 45872\u219445873. Issue 45869 is a different dataset/sample promotion topic and does not match either pair.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue across the full set; the items form two separate duplicate groups, with 45870/45871 and 45872/45873.", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue must be chosen, 45870 is a reasonable anchor because it is an exact duplicate of 45871 and is the earliest of that duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different underlying issue: dataset bundle/sample listing versus an example about financial time series prediction; only broad finance overlap." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different underlying issue: dataset bundle/sample listing versus an example about financial time series prediction; only broad finance overlap." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Exact duplicate titles and matching body size; these look like the same example issue." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Different scope: one is a prediction example, the other is a financial time series plus scraping service; not the same underlying request." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Different scope: one is a prediction example, the other is a financial time series plus scraping service; not the same underlying request." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Different scope: one is a prediction example, the other is a financial time series plus scraping service; not the same underlying request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Different scope: one is a prediction example, the other is a financial time series plus scraping service; not the same underlying request." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Exact duplicate titles and matching body size; these look like the same service-related issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 45870/45871 and 45872/45873 have identical titles and matching body lengths, while 45869 is clearly a different dataset-promotion issue. The soft-edge verdicts are conservative and do not overstate evidence beyond title/body-size duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 20, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46439-2", + "nodes": [ + "issue:46439", + "issue:46441" + ], + "soft_pairs": [ + "issue:46439|issue:46441" + ], + "prepared_review_unit_hash": "3963b456205e92776fb4e21cf71cbd9b50129937c9c4df09037c260dc84a2cc8", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are related in theme (static cache + linear-attention/Qwen3.5), but the evidence is too thin to treat them as the same bug. The observed failures differ (`RuntimeError` from shape corruption vs. `KeyError: 'linear_attention'`), and there are no shared files, targets, or explicit cross-links.", + "confidence": 0.44, + "canonical_issue_reason": "Issue 46439 is the better umbrella report because it describes the suspected mechanism (`early_initialization` corrupting linear-attention layers) more generally, which makes it a stronger anchor for related reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46439 is the best representative for the cluster because it is the more root-cause-oriented writeup, but it should be treated as related-not-duplicate evidence rather than a confirmed duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46439", + "right": "issue:46441", + "accept": false, + "reason": "Only soft topical similarity is present. The bugs surface differently (`RuntimeError` vs. `KeyError: 'linear_attention'`), and there are no shared targets/files or explicit links, so this is not enough to conclude the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The analysis overstates the evidence. The packet only supports a soft similarity between two issue titles; there are no shared files/targets or other concrete links. While both mention static cache and linear-attention/Qwen3.5, the specific failures differ (`RuntimeError` from wrong shapes vs. `KeyError: 'linear_attention'`), so treating them as the same bug is not sufficiently grounded. A more conservative conclusion would be that they may be related, not that they are likely the same issue." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 21, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46551-2", + "nodes": [ + "issue:46551", + "issue:46552" + ], + "soft_pairs": [ + "issue:46551|issue:46552" + ], + "prepared_review_unit_hash": "6b2b5e1f2591b55c58df5c108bf5b7ebe9168baf44d264ba31ccc8fa7c7afa36", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1707, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1707, + "estimated_input_tokens": 427, + "estimated_eval_tokens": 1110 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both issues report the same Wav2Vec2CTCTokenizer v5 bug verbatim; they are duplicates of one another. The open issue is the better canonical record.", + "confidence": 0.99, + "canonical_issue_reason": "issue:46552 is the open copy of the same report and already has inbound references, so it is the better canonical issue to keep.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46552 is the strongest single issue candidate because it is open, identical in content to the other issue, and appears to be the active tracking entry.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46551", + "right": "issue:46552", + "accept": true, + "reason": "The titles and bodies are identical and describe the same tokenizer bug; this is a straightforward duplicate pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded: the packet shows identical titles, identical body lengths, and maximum soft similarity between the two issues. Choosing the open issue with inbound references as the canonical record is conservative and supported. No evidence is overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 22, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46657-2", + "nodes": [ + "issue:46657", + "issue:46692" + ], + "soft_pairs": [ + "issue:46657|issue:46692" + ], + "prepared_review_unit_hash": "a0d1801e51db6111b3ed4402de29422eddbe7a0529c34123232eb740da173991", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1569, + "estimated_input_tokens": 393, + "estimated_eval_tokens": 1042 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1569, + "estimated_input_tokens": 393, + "estimated_eval_tokens": 1042 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates: one reports a tensor-size mismatch during object-detection training, while the other reports out-of-memory after several epochs in RT-DETR v2. They may both occur in training, but they describe different failure modes and likely different root causes.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "46657 is the clearer, more concrete bug report, but it is not the same underlying problem as 46692.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46657", + "right": "issue:46692", + "accept": false, + "reason": "Reject: the reports describe different issues (tensor shape mismatch vs. OOM after epochs) and do not appear to be the same underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 23, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46822-2", + "nodes": [ + "issue:46822", + "issue:46823" + ], + "soft_pairs": [ + "issue:46822|issue:46823" + ], + "prepared_review_unit_hash": "9113889454cfc19309c77b3c85df1a7925c04340d5d3f949d9ffbcd30139bda1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1672, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1672, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate bug reports in different parts of the model-loading stack: one about ZeRO-3 zero.Init failing to shard a composite submodule and causing OOM, the other about 4-bit auto device mapping with max_memory incorrectly sending modules to CPU and raising a ValueError. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the two issues describe different failure modes and different code paths, so neither subsumes the other.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46823 is the clearer self-contained report if a single representative is needed, but it is still not a duplicate of 46822.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46822", + "right": "issue:46823", + "accept": false, + "reason": "Different underlying bugs: ZeRO-3 partitioning/OOM versus 4-bit device_map CPU offload/ValueError. Similar loading context only, not the same fix or code path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 24, + "component_count": 24, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46906-3", + "nodes": [ + "issue:46906", + "issue:46940", + "issue:46948" + ], + "soft_pairs": [ + "issue:46906|issue:46940", + "issue:46906|issue:46948", + "issue:46940|issue:46948" + ], + "prepared_review_unit_hash": "bbf0f291533d1e7cc643658ee1e1c66a62987c5b6ff8089924447d017498c975", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2808, + "estimated_input_tokens": 702, + "estimated_eval_tokens": 1660 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2808, + "estimated_input_tokens": 702, + "estimated_eval_tokens": 1660 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three near-identical \"[serge] integration failure triage\" issues for consecutive dates look like recurring daily triage tickets, not the same underlying bug report. I would keep them separate and avoid duplicate-merging.", + "confidence": 0.87, + "canonical_issue_reason": "issue:46906 is the best anchor: it is the earliest of the three and has the strongest inbound reference count, so it is the most representative triage ticket for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46906 has the most references and earliest timestamp, making it the strongest single issue to point cluster readers to.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46906", + "right": "issue:46940", + "accept": false, + "reason": "Same template/title, but the date-specific triage context suggests separate daily failure reports rather than one shared underlying issue." + }, + { + "left": "issue:46906", + "right": "issue:46948", + "accept": false, + "reason": "These appear to be consecutive-day triage entries; similarity is high, but there is no evidence they describe the exact same bug or fix." + }, + { + "left": "issue:46940", + "right": "issue:46948", + "accept": false, + "reason": "Likely separate daily integration-failure tracking issues, not duplicate reports of one concrete problem." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-2", + "nodes": [ + "pull_request:40546", + "pull_request:42781" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781" + ], + "prepared_review_unit_hash": "1b7f64e9b178bcfb65c1f8832bd58dc865040cef7de327a31a0c5686ce9d60c6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7223, + "estimated_input_tokens": 1806, + "estimated_eval_tokens": 3868 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7223, + "estimated_input_tokens": 1806, + "estimated_eval_tokens": 3868 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs target VibeVoice, but they are not the same concrete change: #40546 is the initial VibeVoice implementation, while #42781 is a draft for VibeVoice Realtime with additional tokenizer/processing and generation changes. Shared files alone are not enough to treat them as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "#40546 is the broader, primary implementation PR with substantial review activity and the initial model integration; #42781 builds on the same model family but is a separate realtime feature branch.", + "best_issue_reason": null, + "best_pr_reason": "#40546 is the best representative of the base VibeVoice integration because it is the main non-draft implementation PR and covers the foundational code path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "They share the VibeVoice subsystem, but #42781 is a distinct realtime add-on with extra generation/processing and acoustic tokenizer work, not the same bug or change as the initial implementation in #40546." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "5892007c7a6f96cc7c7b2f5efc12cd0ef7ddb11d0c91b3a6c8c89e476eabcdf7", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4674, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4674, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. The soft similarity is superficial (both add docs/tests and model support), not the same underlying change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 41224 is the stronger representative artifact: it is a focused, concrete model capability addition with clear implementation scope, docs, and tests.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different models, different tasks, and different code paths; they do not fix the same bug or implement the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41855-13", + "nodes": [ + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048", + "pull_request:46267" + ], + "soft_pairs": [ + "pull_request:41855|pull_request:43823", + "pull_request:41855|pull_request:43424", + "pull_request:42668|pull_request:43823", + "pull_request:43251|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:43424|pull_request:44827", + "pull_request:43424|pull_request:43823", + "pull_request:43823|pull_request:45655", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:46267", + "pull_request:44827|pull_request:46048", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "bec53376d46f663f496be2797675ca0d4641f87f631704f342cd2a645d80baa8", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 18, + "serialized_chars": 34780, + "estimated_input_tokens": 8695, + "estimated_eval_tokens": 17646 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 18, + "serialized_chars": 34780, + "estimated_input_tokens": 8695, + "estimated_eval_tokens": 17646 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly unrelated PRs across different subsystems. The only true near-duplicate is the pair adding kwargs support to `fixed_cross_entropy` in `loss_utils.py` (43251/43254); all other pairs are just thematic or subsystem similarity, not the same underlying change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best canonical PR in the only duplicate pair: it targets the same `fixed_cross_entropy` change, explicitly links the tracked issue, and presents the more complete implementation of forwarding supported kwargs to `nn.functional.cross_entropy`.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest global representative here because it is the clearest duplicate-center candidate and the only PR that cleanly matches another submission of the same fix in the same code path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Different code paths and goals: Mistral tokenizer methods vs adding a new MobileLLM model. Shared \u2018model support\u2019 theme is too broad." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Tokenizer API addition and ExecuTorch dynamic-shape test are unrelated changes; they would not be merged as one PR." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Both touch pretrained loading/processor-related code, but one is a robustness fix for processor auto-loading and the other is a new model addition." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss-function kwargs support and MobileLLM model addition are completely different features." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Different bug fixes in different model families; no shared concrete code-path problem." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "ExecuTorch export tests and Mistral4 test fixes are unrelated test additions, not the same underlying change." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Dynamic-shape export testing is unrelated to adding MobileLLM." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "New model implementation vs class/config resolution ordering fix; no duplicate bug or mergeable shared fix." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "MobileLLM model addition and Falcon output-collection refactor are different model changes with different purposes." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactor and model-creation skill are not the same change; the latter is a meta/tooling PR, not a code-path fix." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon refactor and Mistral4 test fixes touch different models and different problems." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes and CLIP-family `layrnorm` rename are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 test fixes and chat parsing updates are different subsystems and changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 tests and TDT loss kernel integration are unrelated changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 test fixes and `cls.config` resolution ordering are different problems with no shared code path." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Shared auto/modeling files are incidental; one is Mistral4 tests and the other is a model-creation skill draft." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "New model implementation and the model-creation skill are not duplicate code changes." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Same function, same file, same issue target, and both add support for passing cross-entropy kwargs; they are clearly alternative implementations of the same fix and could be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is conservative and grounded in the packet. The only clearly mergeable duplicate pair is PR 43251 / 43254: they touch the same file, same function, and share the same issue target with overlapping `cross_entropy` kwargs support. The other pairwise rejects stay within the evidence shown (mostly only soft similarity with no shared file or issue target), and no overclaiming stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "4b5c38ac0656831c6f2c800a4c003ba2fef06a93eb2223536d9f402f92e39b29", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests are broad dependency bumps in GitHub Actions, but they target different action families and different compatibility goals, so they are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR for duplication here: PR 42975 refreshes multiple GitHub Actions to Node 24-compatible versions, while PR 42976 specifically upgrades docker/build-push-action across Docker-related workflows.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the broader workflow-action maintenance PR and is the closest representative of the cluster, but it still should not subsume PR 42976.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Different underlying changes: 42975 updates several GitHub Actions for Node 24 compatibility, while 42976 only bumps docker/build-push-action in Docker image workflows. They touch overlapping files but not the same concrete bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 6, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-4", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:45490", + "pull_request:46562" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:45490", + "pull_request:45490|pull_request:46562" + ], + "prepared_review_unit_hash": "a7b8f1f88d5bf84d9c5d891b422dca03090c812286368a17c2e4d2fffacb601e", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 13047, + "estimated_input_tokens": 3262, + "estimated_eval_tokens": 6780 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 13047, + "estimated_input_tokens": 3262, + "estimated_eval_tokens": 6780 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are four separate model-addition PRs (ViT NEPA, Molmo, CTSM, Step3p7). They share only broad Hugging Face model-integration boilerplate such as docs and auto-registry updates, not the same concrete code change or bug.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the items are unrelated model integrations rather than duplicates of one another.", + "best_issue_reason": null, + "best_pr_reason": "#46562 is the strongest standalone representative by review activity and scope, but it is still just the Step3p7 model-addition PR, not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different models and different integrations; shared auto/docs files are generic boilerplate, not the same underlying change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "ViT NEPA vs CTSM are unrelated model additions; overlap is limited to registry/docs plumbing." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM are distinct model implementations; similarity is only in common framework registration and docs updates." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46562", + "accept": false, + "reason": "CTSM time-series support and Step3p7 multimodal support are different concrete model changes; shared auto-mapping/docs edits are not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-7", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815", + "pull_request:46380", + "pull_request:46539", + "pull_request:46886", + "pull_request:46922" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:46380", + "pull_request:43291|pull_request:46922", + "pull_request:43665|pull_request:46539", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:46380", + "pull_request:43665|pull_request:46922", + "pull_request:44815|pull_request:46539", + "pull_request:44815|pull_request:46380", + "pull_request:44815|pull_request:46922", + "pull_request:46380|pull_request:46539", + "pull_request:46380|pull_request:46922", + "pull_request:46539|pull_request:46922", + "pull_request:46886|pull_request:46922" + ], + "prepared_review_unit_hash": "f8794db6705d784f5f547f4244f1bcebe0745b7dace2bf6a24e36d4f31ccb5fe", + "component_budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 15, + "serialized_chars": 19604, + "estimated_input_tokens": 4901, + "estimated_eval_tokens": 10058 + }, + "budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 15, + "serialized_chars": 19604, + "estimated_input_tokens": 4901, + "estimated_eval_tokens": 10058 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Mostly unrelated PRs; the only true duplicate pair is the two YOLOS fixes, where PR #46922 subsumes the device-placement correction and the no-split-module guard.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46922 is the best canonical PR because it directly fixes the YOLOS device mismatch in the forward path and also adds the supporting no-split module safeguard for the same bug.", + "best_issue_reason": null, + "best_pr_reason": "PR #46922 is the strongest representative of the cluster: it targets the concrete YOLOS device_map=\"auto\" failure and appears to contain the more complete fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Both are test-only changes in different model areas; they do not describe the same underlying bug or mergeable fix." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper test cleanup vs FP8 dequant/loading fix are unrelated code paths." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "Whisper tokenizer tests and generation watermarking/validation changes are unrelated." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46922", + "accept": false, + "reason": "Whisper test fixes are unrelated to the YOLOS device mismatch fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46539", + "accept": false, + "reason": "These are separate test-only PRs for different models and different failure modes." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "CLIP/ViT test assertions do not match the FP8 dequantization and loading-path change." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "Different subsystems and failure modes: model test cleanup vs generation config validation." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46922", + "accept": false, + "reason": "The test-only CLIP/ViT PR does not fix the YOLOS runtime device mismatch." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46539", + "accept": false, + "reason": "FP8 quantization/loading fixes are unrelated to the Phimoe CI/integration-test environment changes." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "These touch different functionality: dequantization internals vs generation watermarking validation." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46922", + "accept": false, + "reason": "No shared bug: FP8 dequant fix is unrelated to YOLOS device placement." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46539", + "accept": false, + "reason": "Generation-config watermarking validation and Phimoe CI setup are not the same underlying issue." + }, + { + "left": "pull_request:46380", + "right": "pull_request:46922", + "accept": false, + "reason": "Generation config validation changes are unrelated to the YOLOS mismatch bug." + }, + { + "left": "pull_request:46539", + "right": "pull_request:46922", + "accept": false, + "reason": "Phimoe integration-test CI adjustments are unrelated to YOLOS runtime behavior." + }, + { + "left": "pull_request:46886", + "right": "pull_request:46922", + "accept": true, + "reason": "Both PRs target the same YOLOS device-mismatch problem in the same file and same issue target; #46922 adds the actual tensor device move plus the split-module guard, making them mergeable into one fix." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The only accepted duplicate pair, #46886 and #46922, is well-supported by the shared YOLOS file and shared issue target, and the mergeability claim is plausible because #46922 subsumes the device fix plus the no-split-module safeguard. The other pair rejections are also conservative and match the diffs shown." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43451-24", + "nodes": [ + "pull_request:43451", + "pull_request:45144", + "pull_request:46180", + "pull_request:46196", + "pull_request:46658", + "pull_request:46707", + "pull_request:46836", + "pull_request:46850", + "pull_request:46891", + "pull_request:46895", + "pull_request:46902", + "pull_request:46903", + "pull_request:46912", + "pull_request:46920", + "pull_request:46935", + "pull_request:46953", + "pull_request:46954", + "pull_request:46957", + "pull_request:46958", + "pull_request:46961", + "pull_request:46964", + "pull_request:46965", + "pull_request:46966", + "pull_request:46968" + ], + "soft_pairs": [ + "pull_request:46895|pull_request:46958", + "pull_request:46903|pull_request:46958", + "pull_request:46902|pull_request:46965", + "pull_request:46902|pull_request:46957", + "pull_request:46954|pull_request:46965", + "pull_request:46954|pull_request:46961", + "pull_request:46180|pull_request:46658", + "pull_request:46658|pull_request:46903", + "pull_request:46891|pull_request:46964", + "pull_request:45144|pull_request:46912", + "pull_request:46891|pull_request:46954", + "pull_request:43451|pull_request:46954", + "pull_request:46954|pull_request:46957", + "pull_request:46903|pull_request:46964", + "pull_request:46964|pull_request:46966", + "pull_request:46196|pull_request:46902", + "pull_request:46658|pull_request:46902", + "pull_request:43451|pull_request:46902", + "pull_request:46935|pull_request:46957", + "pull_request:46958|pull_request:46961", + "pull_request:45144|pull_request:46196", + "pull_request:43451|pull_request:45144", + "pull_request:46891|pull_request:46935", + "pull_request:43451|pull_request:46964", + "pull_request:43451|pull_request:46935", + "pull_request:45144|pull_request:46707", + "pull_request:46935|pull_request:46961", + "pull_request:45144|pull_request:46658", + "pull_request:46935|pull_request:46953", + "pull_request:46707|pull_request:46920", + "pull_request:46912|pull_request:46920", + "pull_request:46658|pull_request:46964", + "pull_request:46836|pull_request:46964", + "pull_request:46196|pull_request:46964", + "pull_request:46958|pull_request:46965", + "pull_request:43451|pull_request:46920", + "pull_request:46920|pull_request:46957", + "pull_request:46957|pull_request:46958", + "pull_request:46836|pull_request:46920", + "pull_request:46891|pull_request:46958", + "pull_request:43451|pull_request:46836", + "pull_request:46891|pull_request:46957", + "pull_request:46658|pull_request:46836", + "pull_request:43451|pull_request:46658", + "pull_request:43451|pull_request:46196", + "pull_request:46953|pull_request:46957", + "pull_request:46707|pull_request:46953", + "pull_request:43451|pull_request:46891", + "pull_request:46658|pull_request:46850", + "pull_request:46891|pull_request:46953", + "pull_request:46658|pull_request:46707", + "pull_request:46707|pull_request:46891", + "pull_request:43451|pull_request:46707", + "pull_request:46836|pull_request:46891", + "pull_request:43451|pull_request:46953", + "pull_request:46836|pull_request:46957", + "pull_request:46658|pull_request:46957", + "pull_request:46961|pull_request:46965", + "pull_request:46707|pull_request:46836", + "pull_request:46957|pull_request:46965", + "pull_request:46953|pull_request:46968", + "pull_request:46707|pull_request:46957", + "pull_request:46196|pull_request:46957", + "pull_request:46912|pull_request:46957", + "pull_request:46957|pull_request:46968", + "pull_request:46850|pull_request:46957", + "pull_request:43451|pull_request:46912", + "pull_request:46850|pull_request:46891", + "pull_request:46891|pull_request:46968", + "pull_request:43451|pull_request:46957", + "pull_request:46196|pull_request:46836", + "pull_request:43451|pull_request:46850" + ], + "prepared_review_unit_hash": "55caed96e7fac2c508287592d92adabf7a36a4c3ba90a9e4541972b679923bbf", + "component_budget": { + "node_count": 24, + "item_count": 24, + "soft_pair_count": 86, + "serialized_chars": 90472, + "estimated_input_tokens": 22618, + "estimated_eval_tokens": 45492 + }, + "budget": { + "node_count": 24, + "item_count": 24, + "soft_pair_count": 72, + "serialized_chars": 79405, + "estimated_input_tokens": 19852, + "estimated_eval_tokens": 39960 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated PRs that happen to look similar at a text/surface level. The only clear duplicate pair is the CohereASR training-loss double-shift fix (46895 and 46958); the rest are different model additions, distinct bug fixes, CI/docs changes, or separate infrastructure work.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46895 is the best canonical PR because it is the clearest representative of the one genuine duplicate pair: it fixes the same CohereASR training-loss double-shift bug as 46958, targets the same issue, and has the stronger review history.", + "best_issue_reason": null, + "best_pr_reason": "PR 46895 is the strongest overall representative: it is the only clearly duplicated change in the set and is the more established version of that CohereASR fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46895", + "right": "pull_request:46958", + "accept": true, + "reason": "Same CohereASR training-loss double-shift bug in the same code path/files; 46958 is an alternate implementation of the same fix." + }, + { + "left": "pull_request:46903", + "right": "pull_request:46958", + "accept": false, + "reason": "Both are training-loss fixes, but for different models (PPFormulaNet vs CohereASR); not the same underlying change." + }, + { + "left": "pull_request:46902", + "right": "pull_request:46965", + "accept": false, + "reason": "Offloading/save-state changes vs DeepSeek V4 cache handling; unrelated bug fixes." + }, + { + "left": "pull_request:46902", + "right": "pull_request:46957", + "accept": false, + "reason": "Different areas: disk-offloaded buffer saving vs generation decode optimization." + }, + { + "left": "pull_request:46954", + "right": "pull_request:46965", + "accept": false, + "reason": "Different model internals: FusedRMSNormGated dtype cleanup vs DeepSeek V4 use_cache behavior." + }, + { + "left": "pull_request:46954", + "right": "pull_request:46961", + "accept": false, + "reason": "A code clean-up in model modules vs a workflow branch-normalization PR; unrelated." + }, + { + "left": "pull_request:46180", + "right": "pull_request:46658", + "accept": false, + "reason": "Different model additions (Fun-ASR-Nano vs Step-3.7-Flash), not the same change." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46903", + "accept": false, + "reason": "Step-3.7-Flash model build/conversion work vs PPFormulaNet loss fix; unrelated." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46964", + "accept": false, + "reason": "CI/workflow changes vs quantized-PEFT training validation; different purpose and code path." + }, + { + "left": "pull_request:45144", + "right": "pull_request:46912", + "accept": false, + "reason": "MiMo-V2 model addition vs RoPE test expansion; not the same underlying bug/change." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46954", + "accept": false, + "reason": "CI integration changes vs model dtype clean-up; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46954", + "accept": false, + "reason": "Molmo2 model addition vs FusedRMSNormGated dtype change; different work." + }, + { + "left": "pull_request:46954", + "right": "pull_request:46957", + "accept": false, + "reason": "Model dtype/module cleanup vs generation kernel optimization; unrelated." + }, + { + "left": "pull_request:46903", + "right": "pull_request:46964", + "accept": false, + "reason": "Different fixes in different models; only share a broad training context." + }, + { + "left": "pull_request:46964", + "right": "pull_request:46966", + "accept": false, + "reason": "Quantization validation vs Gemma4Unified sequence-classification support; different feature areas." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46902", + "accept": false, + "reason": "SDPA attention docs/code vs offloading/save conversion; unrelated." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46902", + "accept": false, + "reason": "Step-3.7-Flash model work vs offloading/save-state bugfix; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46902", + "accept": false, + "reason": "Molmo2 model addition vs offloading/save-state bugfix; unrelated." + }, + { + "left": "pull_request:46935", + "right": "pull_request:46957", + "accept": false, + "reason": "Trackio callback finalization vs generation decode optimization; different subsystems." + }, + { + "left": "pull_request:46958", + "right": "pull_request:46961", + "accept": false, + "reason": "CohereASR loss fix vs Serge workflow normalization; unrelated." + }, + { + "left": "pull_request:45144", + "right": "pull_request:46196", + "accept": false, + "reason": "Different model additions/fixes (MiMo-V2 vs wav2vec2 SDPA dispatch)." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46891", + "accept": false, + "reason": "Model addition vs CI workflow change; not the same change." + }, + { + "left": "pull_request:45144", + "right": "pull_request:46707", + "accept": false, + "reason": "Model addition vs FSDP2 integration work; unrelated." + }, + { + "left": "pull_request:46935", + "right": "pull_request:46961", + "accept": false, + "reason": "Trackio callback fix vs workflow normalization; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46658", + "accept": false, + "reason": "Different new-model additions (Molmo2 vs Step-3.7-Flash), not duplicates." + }, + { + "left": "pull_request:46920", + "right": "pull_request:46957", + "accept": false, + "reason": "mistral-common API adaptation for Voxtral Realtime vs generation decode optimization; unrelated." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46958", + "accept": false, + "reason": "One is generation optimization, the other is a CohereASR loss bugfix; different code paths." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46920", + "accept": false, + "reason": "Unlimited OCR model addition vs Voxtral Realtime API adaptation; unrelated." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46958", + "accept": false, + "reason": "CI/workflow work vs CohereASR training-loss fix; different changes." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46965", + "accept": false, + "reason": "Generation optimization vs DeepSeek V4 cache semantics; not the same bug." + }, + { + "left": "pull_request:46953", + "right": "pull_request:46968", + "accept": false, + "reason": "Model dtype cleanup vs docs autolink fixes; unrelated." + }, + { + "left": "pull_request:46707", + "right": "pull_request:46957", + "accept": false, + "reason": "FSDP2 migration vs generation optimization; different infrastructure areas." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46957", + "accept": false, + "reason": "Wav2vec2 SDPA support vs generation optimization; unrelated." + }, + { + "left": "pull_request:46912", + "right": "pull_request:46957", + "accept": false, + "reason": "RoPE testing improvements vs generation optimization; different intent and code." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46968", + "accept": false, + "reason": "Generation optimization vs docs autolink cleanup; unrelated." + }, + { + "left": "pull_request:46850", + "right": "pull_request:46957", + "accept": false, + "reason": "Gemma sliding-window attention masking bugfix vs generation optimization; different underlying issue." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46912", + "accept": false, + "reason": "Molmo2 model addition vs RoPE tests; unrelated." + }, + { + "left": "pull_request:46850", + "right": "pull_request:46891", + "accept": false, + "reason": "Gemma masking bugfix vs CI workflow changes; unrelated." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46968", + "accept": false, + "reason": "CI workflow change vs docs autolink cleanup; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46957", + "accept": false, + "reason": "Molmo2 model addition vs generation optimization; unrelated." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46836", + "accept": false, + "reason": "Wav2vec2 SDPA dispatch vs Unlimited OCR model addition; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46836", + "accept": false, + "reason": "Different model additions (Molmo2 vs Unlimited OCR); not the same underlying change." + }, + { + "left": "pull_request:46658", + "right": "pull_request:46836", + "accept": false, + "reason": "Step-3.7-Flash model build vs Unlimited OCR model addition; unrelated." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46196", + "accept": false, + "reason": "Molmo2 model addition vs wav2vec2 SDPA dispatch; unrelated." + }, + { + "left": "pull_request:46953", + "right": "pull_request:46957", + "accept": false, + "reason": "Model dtype cleanup vs generation optimization; unrelated." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46957", + "accept": false, + "reason": "CI integration work vs generation optimization; unrelated." + }, + { + "left": "pull_request:46707", + "right": "pull_request:46836", + "accept": false, + "reason": "FSDP2 migration vs Unlimited OCR model addition; unrelated." + }, + { + "left": "pull_request:46961", + "right": "pull_request:46965", + "accept": false, + "reason": "Workflow normalization vs DeepSeek V4 cache support; unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted duplicate pair is the CohereASR training-loss double-shift fix (46895/46958), which is supported by the shared files and shared issue target. The other verdicts appropriately treat surface similarity as insufficient evidence for duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 36, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43451-24", + "nodes": [ + "pull_request:43451", + "pull_request:46196", + "pull_request:46707", + "pull_request:46836", + "pull_request:46850", + "pull_request:46891", + "pull_request:46912", + "pull_request:46953", + "pull_request:46957", + "pull_request:46961", + "pull_request:46965", + "pull_request:46968" + ], + "soft_pairs": [ + "pull_request:46707|pull_request:46850", + "pull_request:46891|pull_request:46912", + "pull_request:46836|pull_request:46912", + "pull_request:43451|pull_request:46968", + "pull_request:46707|pull_request:46912", + "pull_request:46707|pull_request:46968", + "pull_request:43451|pull_request:46965", + "pull_request:46196|pull_request:46953", + "pull_request:46836|pull_request:46965", + "pull_request:46196|pull_request:46891", + "pull_request:46957|pull_request:46961", + "pull_request:46891|pull_request:46961", + "pull_request:43451|pull_request:46961", + "pull_request:46953|pull_request:46961" + ], + "prepared_review_unit_hash": "f0094e2801060f78c9e1babd3c06b844f41dddb046a7abbe209951a25661612f", + "component_budget": { + "node_count": 24, + "item_count": 24, + "soft_pair_count": 86, + "serialized_chars": 90472, + "estimated_input_tokens": 22618, + "estimated_eval_tokens": 45492 + }, + "budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 14, + "serialized_chars": 33353, + "estimated_input_tokens": 8339, + "estimated_eval_tokens": 16934 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: model additions, unrelated bug fixes, docs, CI/workflow changes, and performance/infrastructure work. I found no true duplicate pair among the soft edges, so no single canonical PR stands out.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs do not share the same underlying bug or change; they span unrelated features and fixes across models, docs, CI, and distributed/runtime code.", + "best_issue_reason": null, + "best_pr_reason": "No global best PR for deduplication: none of these PRs is a safe representative for the cluster because they are not semantically overlapping enough to merge or substitute for one another.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46707", + "right": "pull_request:46850", + "accept": false, + "reason": "FSDP2 module/migration work vs Gemma 3/4 sliding-window attention masking; different code paths and different problems." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46912", + "accept": false, + "reason": "CI workflow for PEFT integration vs RoPE testing improvements; both are infra/test related but not the same change." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46912", + "accept": false, + "reason": "Unlimited OCR model addition vs RoPE test coverage; unrelated model and test work." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46968", + "accept": false, + "reason": "Molmo2 model addition vs docs autolink cleanup; unrelated feature vs documentation fix." + }, + { + "left": "pull_request:46707", + "right": "pull_request:46912", + "accept": false, + "reason": "Native FSDP2 migration vs RoPE test expansion; no shared concrete bug or code path." + }, + { + "left": "pull_request:46707", + "right": "pull_request:46968", + "accept": false, + "reason": "FSDP2 runtime/distributed changes vs docs autolink cleanup; entirely different scope." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46965", + "accept": false, + "reason": "Molmo2 model add vs DeepSeek V4 use_cache fix; different models and issues." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46953", + "accept": false, + "reason": "wav2vec2 SDPA dispatch support vs removing default dtype in FusedRMSNormGated modules; unrelated model/runtime changes." + }, + { + "left": "pull_request:46836", + "right": "pull_request:46965", + "accept": false, + "reason": "Unlimited OCR model addition vs DeepSeek V4 cache handling; different models and distinct fixes." + }, + { + "left": "pull_request:46196", + "right": "pull_request:46891", + "accept": false, + "reason": "wav2vec2 SDPA dispatch/model changes vs PEFT CI workflow; not the same underlying change." + }, + { + "left": "pull_request:46957", + "right": "pull_request:46961", + "accept": false, + "reason": "FP8 generation kernel optimization vs Serge branch-normalization workflow; unrelated changes." + }, + { + "left": "pull_request:46891", + "right": "pull_request:46961", + "accept": false, + "reason": "PEFT integration CI work vs Serge fix normalization workflow; both touch CI but solve different problems." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46961", + "accept": false, + "reason": "Molmo2 model addition vs automated patch-normalization workflow; unrelated." + }, + { + "left": "pull_request:46953", + "right": "pull_request:46961", + "accept": false, + "reason": "FusedRMSNormGated dtype cleanup vs Serge normalization workflow; no shared bug or mergeable code path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-3", + "nodes": [ + "pull_request:43488", + "pull_request:43532", + "pull_request:46747" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:46747", + "pull_request:43488|pull_request:43532", + "pull_request:43532|pull_request:46747" + ], + "prepared_review_unit_hash": "3c346fea6026ace24b0ef8113944e7f5873423c4c5ba8c735d60388c57c89522", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 6350, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 6350, + "estimated_input_tokens": 1588, + "estimated_eval_tokens": 3432 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three PRs are all synthetic '[don't merge]' changes, but they do not appear to fix the same underlying problem. They span unrelated areas: packaging/versioning, PR-comment/workflow automation, and a security-gate workflow change.", + "confidence": 0.36, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43532 is the broadest and most representative of the CI/testing automation changes in this set, touching multiple workflow and test-related files.", + "best_issue_reason": null, + "best_pr_reason": "PR #43532 is the strongest standalone representative because it has the widest scope and the most context across workflow and testing utilities.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:46747", + "accept": false, + "reason": "One is a packaging/versioning and test-dependency change; the other only flips a workflow trigger for the security gate. Different code paths and different fixes." + }, + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "They overlap only on a Vit test file, but the actual changes are unrelated: one is setup/dependency/versioning, the other is workflow automation plus test helpers." + }, + { + "left": "pull_request:43532", + "right": "pull_request:46747", + "accept": false, + "reason": "Both touch workflows, but they address different CI paths and different behavior: self-comment/slack/scheduled automation versus the PR security gate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "bccfaddb20abd42e853295edafbe3b6ee6eee267825d3a6a3a6dd01a240f47d7", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate PRs adding different segmentation pipelines: concept segmentation vs visual segmentation. They share scaffolding and docs, but target different tasks and code paths, so they should not be deduplicated.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different features and implementations: Promptable Concept Segmentation vs Promptable Visual Segmentation. Shared docs/pipeline boilerplate is not enough to treat them as the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "d6196c36fa4e3b9207533e0a7b2683ebaedb7be0e7942e6afb2d2ce1143f68b5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs touch `src/transformers/trainer.py`, but they address different trainer extensibility points: gradient-accumulation loss scaling vs. overriding signature-column filtering. They are related in theme, not the same underlying change or bug.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43651 is the cleaner representative of the cluster because it introduces a concrete behavior fix in `training_step` with a narrowly scoped API hook; PR 43663 is a separate refactor/extension for column filtering.", + "best_issue_reason": null, + "best_pr_reason": "PR 43651 is the best single representative: it has the more direct bug-fix/change semantics and a self-contained code-path impact, while 43663 is an independent trainer customization change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different code paths and problems: 43651 changes loss scaling during gradient accumulation; 43663 changes how trainer signature columns are derived for filtering. Shared file/subsystem alone is not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-2", + "nodes": [ + "pull_request:43656", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842" + ], + "prepared_review_unit_hash": "1e9b7f4b82b2b4abaaff88e99d9d07b319b0aa2de447ad3873500d10bc056154", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs touch the same CLI file and both mention TypeAdapter, but they fix different NameError causes and are not safe duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the more specific, concrete fix: it addresses the runtime NameError when pydantic is absent by guarding the TypeAdapter annotation/import. PR 43656 is a broader typing cleanup around annotations in the same file.", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 is the best representative of the cluster because it targets a clear runtime failure mode in the CLI. PR 43656 is related but solves a different annotation-related problem.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Same file and same symbol, but different bugs: one is a general annotation/type-hint NameError fix, the other specifically prevents importing TypeAdapter when pydantic is missing. They are not the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43779-2", + "nodes": [ + "pull_request:43779", + "pull_request:43816" + ], + "soft_pairs": [ + "pull_request:43779|pull_request:43816" + ], + "prepared_review_unit_hash": "e7c674adcc031aa8883109a5dfc678b4407fb20b10764302092e384d031ba83a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests appear to implement the same SwanLab integration change: adding run id/resume support in `integration_utils.py` for issue #43698, with one variant also expanding the docs. They are strong duplicate PR candidates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43816 is the cleaner representative: it carries the same SwanLab id/resume support plus the accompanying documentation updates, making it the more complete version of the shared change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43816 is the best global representative because it appears to include the full feature change and docs for SwanLab run id/resume support, while #43779 is the narrower implementation-focused variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43779", + "right": "pull_request:43816", + "accept": true, + "reason": "Same underlying SwanLab callback enhancement for id/resume support in the same file and same issue target; the doc differences do not change the core code-path change, so these are duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target issue #43698, touch the same file, and the previews indicate the same SwanLab id/resume support with #43816 adding docs. The duplicate/mergeable verdict is conservative enough given the shared target and overlapping code change." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44722", + "pull_request:44018|pull_request:44116", + "pull_request:44068|pull_request:44116", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44085", + "pull_request:44018|pull_request:44086", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44085|pull_request:44086", + "pull_request:44059|pull_request:44086", + "pull_request:44098|pull_request:44101", + "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44722", + "pull_request:44017|pull_request:44073", + "pull_request:44068|pull_request:44071", + "pull_request:44071|pull_request:44086", + "pull_request:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44073|pull_request:44101", + "pull_request:44073|pull_request:44161", + "pull_request:44059|pull_request:44071", + "pull_request:44056|pull_request:44101", + "pull_request:44018|pull_request:44059", + "pull_request:44059|pull_request:44068", + "pull_request:44101|pull_request:44161", + "pull_request:43997|pull_request:44017", + "pull_request:43997|pull_request:44056", + "pull_request:44071|pull_request:44076", + "pull_request:44074|pull_request:44101", + "pull_request:43997|pull_request:44018", + "pull_request:44056|pull_request:44161", + "pull_request:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:44018|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", + "pull_request:43997|pull_request:44074", + "pull_request:44010|pull_request:44018", + "pull_request:44071|pull_request:44129", + "pull_request:43997|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154" + ], + "prepared_review_unit_hash": "67531753cef017df60c4b6a2937f1adbc6cb8813a8a79374976431a7dccb3bce", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75098, + "estimated_input_tokens": 18775, + "estimated_eval_tokens": 37806 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is one umbrella output-tracing refactor campaign tied to issue #43979, with a few true duplicates where PRs touch the same concrete model code path (especially GPT-J / CodeGen and GPT-Neo variants), but most pairs are just same-pattern refactors across different models and should stay separate.", + "confidence": 0.95, + "canonical_issue_reason": "#43979 is the shared umbrella tracking issue referenced by the refactor PRs and best represents the common output-tracing work.", + "canonical_pr_reason": "PR #44722 is the strongest canonical representative: it covers the same GPT-J/CodeGen output-tracing refactor as the overlapping PRs and appears to be the most complete/latest version of that concrete change.", + "best_issue_reason": "#43979 is the only issue-level umbrella in the cluster and is the natural anchor for the duplicated refactor PRs.", + "best_pr_reason": "#44722 is the best overall PR representative for the duplicated GPT-J/CodeGen refactor slice of this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models and code paths; same issue style only." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different model families; not the same concrete bug/change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both target GPT-J output tracing on the same code path; 44066 is just a broader variant that also includes CodeGen." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "Both are output-tracing refactors, but for different vision backbones and different implementations." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet vs GPT-J/CodeGen; unrelated concrete changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "Different models and forward paths." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa are unrelated code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen refactor vs MGP-STR refactor; different model code." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "Different model architectures and forward implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both are GPT-Neo output-tracing refactors touching the same model code path and look mergeable as one PR." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen versus GPT-Neo; same theme, not the same change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "Different models and different concrete code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet refactor and GPT-J/CodeGen refactor are unrelated changes." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are different model paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated concrete implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different model code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet vs MPT; not the same underlying bug/change." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo refactor is separate from GPT-J/CodeGen refactor." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated implementations." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are different code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT refactor is not the same concrete change as GPT-J/CodeGen." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo versus GPT-J; same refactor pattern but different model code." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different model paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated concrete implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are unrelated model changes." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are unrelated changes." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 refactor is separate from GPT-J/CodeGen." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are unrelated implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer refactor is separate from GPT-J/CodeGen." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are different model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are unrelated concrete changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are different model implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are unrelated refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are different model code paths." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are unrelated changes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are different model implementations." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are unrelated concrete refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44161", + "accept": false, + "reason": "GPT-Neo and LongT5 are different code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are unrelated model changes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are different code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are unrelated concrete changes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are different model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "SqueezeBert and GPT-Neo are unrelated code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTa v2 are different concrete changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa v2 and TextNet are unrelated refactors." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen output-tracing refactor on the same concrete files; clearly duplicate territory." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same model path and could be merged into one PR." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are unrelated changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "FNet/CVT versus ResNet/RegNet/RT-DETR; not the same bug or change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "FNet/CVT and SpeechT5 are different code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa v2 and SpeechT5 are unrelated model changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "FNet/CVT and SegFormer are different implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "FNet/CVT and LongT5 are unrelated refactors." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen and EfficientNet are different model code paths." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa v2 are unrelated." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen and MGP-STR are different model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are unrelated concrete changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Same GPT-Neo output-tracing refactor; same code path and mergeable as one PR." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen versus GPT-Neo; same pattern, different model code." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are distinct model paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet refactor is unrelated to GPT-J/CodeGen." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are different concrete changes." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different model code paths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are unrelated changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo is separate from GPT-J/CodeGen." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are different concrete refactors." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are different model paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J are distinct refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different code paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated model implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different changes." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are unrelated." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 refactor is unrelated to GPT-J/CodeGen." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are unrelated refactors." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J/CodeGen are unrelated." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are different model implementations." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are unrelated concrete changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are different model paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are unrelated changes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are different model code paths." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are unrelated refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are different concrete changes." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are unrelated." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44161", + "accept": false, + "reason": "GPT-Neo and LongT5 are different model paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are unrelated model changes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are different code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are unrelated refactors." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are different model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "SqueezeBert and GPT-Neo are unrelated changes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTa v2 are different concrete code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa v2 and TextNet are unrelated." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa v2 and VITS are different model code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded in the packet: #43979 is the clear umbrella issue, and #44722 is a reasonable canonical PR for the GPT-J/CodeGen slice because it is the latest/most complete overlapping change. The soft-edge calls are conservative overall: the only true-duplicate verdicts are for the same model-path refactors (GPT-J/CodeGen and GPT-Neo), while the rest are appropriately separated by model/code path." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 36, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44024", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44024|pull_request:44076", + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "4cadede3443298b9f5cd151b2903f29c2ace5152c0423413cee974d44059f093", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84902, + "estimated_input_tokens": 21226, + "estimated_eval_tokens": 42708 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17125, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a family of model-specific refactors to the standardized output-capturing / tracing APIs (capture_outputs, can_return_tuple, _can_record_outputs). They are thematically similar but mostly independent changes across different model implementations, not duplicate fixes for one underlying bug.", + "confidence": 0.74, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44129 is the best anchor for the cluster: it has an explicit issue target (43979) and a concrete SpeechT5 output-tracing refactor, making it the most clearly grounded representative change.", + "best_issue_reason": null, + "best_pr_reason": "PR 44129 is the strongest representative of the shared migration theme because it is explicitly tied to an issue and updates a concrete code path rather than just following the broader refactor pattern.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are output-tracing migrations, but they touch different model code paths (FocalNet vs ImageGPT) with model-specific hooks and decorator wiring. Same general refactor theme, not the same underlying bug/change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBert and TextNet are separate model implementations. The diffs show similar decorator migration work, but no shared concrete fix or mergeable single PR scope." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "These are parallel output-capture refactors in different models (MPNet vs VisualBert). They are related by infrastructure, not by a single bug or code-path correction." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "MPNet and TextNet are distinct model-specific refactors; the overlap is only the standardized tracing pattern, which is too broad to treat as duplicate work." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert changes are both migration-style edits, but they affect different model internals and do not appear to fix the same concrete defect." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 are unrelated model code paths. Shared use of output-capturing utilities is insufficient to merge them as one change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44028", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030" + ], + "prepared_review_unit_hash": "88de7bc6d7ab46b44b67dbd9179e353d1bb79c54e375c965fc62f99279ee84e8", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40183, + "estimated_input_tokens": 10046, + "estimated_eval_tokens": 20348 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40183, + "estimated_input_tokens": 10046, + "estimated_eval_tokens": 20348 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This is a loose cluster of independent, model-specific \"refactor output tracing\" pull requests. They share a common refactoring theme, but they target different architectures and code paths, so they should not be merged as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44003 is the most representative of the cluster: it makes the broadest output-tracing change, spans two related model files, and touches the shared output-capturing plumbing rather than only a single wrapper.", + "best_issue_reason": null, + "best_pr_reason": "PR 44003 is the strongest global representative because it is the most infrastructure-level output-tracing refactor in the set and is broader than the single-model adaptations.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Same refactor theme, but `timm_backbone` and `upernet` are unrelated model paths with different output logic." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Both mention output tracing, but these are different model families and different concrete changes." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Unrelated architectures (`timm_backbone` vs `univnet`); not the same underlying bug or change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Different code paths and mechanisms; one is a backbone refactor, the other is mamba output-capture plumbing." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and different forward/output semantics; too broad to be duplicates." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "Both are output-tracing refactors, but they affect different model implementations and are not mergeable as one PR." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model code paths (`mobilenet_v1` vs `upernet`), so this is only thematic similarity." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Separate model families with distinct tracing changes; not the same concrete fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "`mobilenet_v1` and `vision_encoder_decoder` change different forward APIs and output handling." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Different encoder-decoder vs CNN model code paths; only the refactor theme overlaps." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "`vision_text_dual_encoder` and `univnet` are unrelated; the output-tracing changes are model-specific." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model subsystems and different output shapes/return conventions." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "These touch unrelated architectures and different tracing implementations, not one shared bug." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and separate output-tracing refactors; not the same concrete change." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models (`univnet` vs `upernet`) and different forward signatures; too broad to merge." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "`univnet` and `mamba` fix different output-tracing paths in unrelated code." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Separate model families; same refactor motif is not enough for duplicate triage." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "`upernet` and `rwkv` are different architectures with unrelated output handling." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model stacks and different implementation details; not the same underlying change." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "`mamba`/`falcon_mamba` output-capture plumbing is unrelated to `superpoint`'s refactor." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "`codegen` changes cache/attention forward typing; this is a different code-path change than mamba tracing." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model families and different refactor targets; not the same bug or feature." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "`univnet` and `codegen` are unrelated, and the diffs alter different forward semantics." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "`upernet` and `codegen` have unrelated output-tracing/caching changes." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Different architectures and different concrete output-path changes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Both are tracing refactors, but they target different model families (`depth_anything` vs `vision_encoder_decoder`)." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model families and different forward/output conventions." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "`depth_anything` and `superpoint` are separate model implementations with distinct changes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Different architectures and different tracing mechanisms; not mergeable as one PR." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "`depth_anything` and `dpr` modify unrelated model outputs and signatures." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "`vision_encoder_decoder` and `rwkv` are unrelated subsystems with different output handling." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Different encoder-decoder wrappers for different modalities; not the same change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "`vision_encoder_decoder` vs `superpoint` are different code paths and output formats." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "These touch different model families and different wrapper behavior." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "Speech encoder-decoder and SuperPoint have unrelated forward/output refactors." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different models and different output-tracing mechanisms." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "`speech_encoder_decoder` and `dpr` are unrelated code paths." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "`superpoint` and `rwkv` are different architectures; the changes are only superficially similar." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model implementations and different output semantics." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "`rwkv` output tracing and `dpr` encoder wrapping are unrelated concrete changes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-2", + "nodes": [ + "pull_request:44662", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401" + ], + "prepared_review_unit_hash": "da3d2c27c3d69ed13c3b89eb712dc1eecd11e84e7d910c7f01770742d5d27ed9", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate model-implementation PRs for different architectures (PenguinVL vs Voxtral TTS), not duplicates. They share some common Transformers plumbing and docs, but they fix/add different concrete code paths.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45401 is the cleaner representative of the cluster only in the sense that it is more focused and explicitly tied to issue 45027, but it is not a duplicate of 44662.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different model additions with different purposes and code paths: PenguinVL implementation vs Voxtral TTS support. Shared auto/docs files are generic registration work, not evidence of the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44965-2", + "nodes": [ + "pull_request:44965", + "pull_request:46845" + ], + "soft_pairs": [ + "pull_request:44965|pull_request:46845" + ], + "prepared_review_unit_hash": "9fe18532c9b7e6bbf94eb214d0b71edb108805a69147dc5b802423e1e2d8bd2b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4218, + "estimated_input_tokens": 1055, + "estimated_eval_tokens": 2366 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4218, + "estimated_input_tokens": 1055, + "estimated_eval_tokens": 2366 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are not duplicates: one is a draft/CI workflow experiment to convert PRs to draft, while the other changes CI branch routing and includes an unrelated placeholder code edit. They do not appear to fix the same concrete problem.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "No clear canonical PR: #44965 is a malformed workflow experiment, while #46845 is a separate CI branch-selection tweak plus an unrelated placeholder code change; they are not the same change.", + "best_issue_reason": null, + "best_pr_reason": "#46845 is the stronger standalone candidate because it makes a concrete CI routing change, whereas #44965 looks like an incomplete workflow experiment with no clear finished fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44965", + "right": "pull_request:46845", + "accept": false, + "reason": "Reject: the PRs target different behaviors. #44965 aims to convert PRs to draft via a workflow rewrite, while #46845 changes CI branch handling and includes an unrelated code placeholder edit; they are not the same underlying bug or mergeable fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "43cfed12d1b60f998b66bd65d5de06bc081c4afdd54a20f85ff3b4352c28aebc", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The PRs are related but not duplicates: one introduces a generic module-fusion API, while the other integrates KernelConfig-based n-to-1 kernel fusion into model loading and adds an example.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44979 is the broader foundational change: it introduces the generic module_fusion API and tests. PR 45363 is a narrower, draft integration on top of that idea, so 44979 is the better representative of the underlying work.", + "best_issue_reason": null, + "best_pr_reason": "PR 44979 is the strongest representative in this cluster because it is the core, non-draft API change and is broader than the specific KernelConfig fusion integration in 45363.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Related feature area, but not the same concrete change. 44979 adds a new generic module-fusion API; 45363 wires KernelConfig into loading for a specific n-to-1 fusion path and includes an example. They could be merged separately." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "69a94bdf6f2dd83ae97ac7b128e9525d88c66c4d587c8de87b9c0033f5d3a574", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs patch the same Mistral regex pre-tokenizer AttributeError in the same file and on the same lines; 45317 is the more complete version because it adds tests as well as the code fix.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "45317 is the better canonical PR: it fixes the same bug as 45086 but also includes test coverage, making it the more complete representative of the change.", + "best_issue_reason": null, + "best_pr_reason": "45317 is the strongest PR in the cluster because it addresses the concrete AttributeError and adds validation tests, while 45086 is a smaller precursor of the same fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying code-path bug in _patch_mistral_regex, same target issue, and the diffs overlap directly; the later PR plausibly subsumes the earlier one." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "0153660c376a5927927682305f06a1ec4241b6c9b761ee3c01fadc2310693163", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs overlap on NemotronH internals, but they are not the same change: one is a refactor toward GraniteMoeHybrid-style inheritance/conversion mapping, while the other is a broader model split introducing NemotronHDense/NemotronHSparse with docs and tests. Treat as related, not duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true canonical duplicate here; if one must be chosen as the broader representative, PR #45569 is the more complete user-facing model change, but it is not the same underlying fix as #45115.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the strongest standalone PR in the cluster because it includes the architecture split, model registrations, docs, and tests. PR #45115 is a narrower refactor/conversion mapping change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Shared NemotronH files are not enough: #45115 is a focused refactor/conversion mapping/inheritance change, while #45569 is a broader NemotronH rework that adds dense/sparse variants and documentation/tests. They do not look mergeable into one duplicate PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "bbe087218a903ef8449c8967779824ff1adc877c61632828d0787c5018b06db1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5953, + "estimated_input_tokens": 1489, + "estimated_eval_tokens": 3234 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5953, + "estimated_input_tokens": 1489, + "estimated_eval_tokens": 3234 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two related but distinct open PRs: one adds full heterogeneous model support including modeling/runtime changes, while the other is a narrower config-only subset. They overlap heavily in docs/config files but are not the same concrete change.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45332 is the better canonical representative because it is the broader umbrella change and includes the config work plus the downstream modeling/runtime pieces. PR 45333 is a narrower config-only slice.", + "best_issue_reason": null, + "best_pr_reason": "PR 45332 is the best single PR for the cluster since it captures the full heterogeneous model feature and subsumes the shared configuration work.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Related feature work, but not the same underlying change: 45332 spans heterogeneous modeling/runtime support, while 45333 focuses only on per-layer configuration. They would not plausibly be merged as one duplicate PR without losing scope distinctions." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "4f6620af032f973cb13c41d51ff29b87e2230985bbd8622d340405e59f4730df", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests are typing-related, but they address different kinds of work: one is a broad type-checking sweep across many transformer modules, while the other adds specific typing protocols/aliases and related fixes. They are similar in theme, but not the same underlying change.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the more focused and concrete typing change set; PR 45415 is a larger sweep of type-checking adjustments and unrelated file-by-file cleanups, so it is less representative of a single duplicate target.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Shared files and a typing theme are not enough here: 45415 is a broad mypy/type-checking pass, while 45425 introduces specific typing aliases/protocols and targeted fixes. They do not look like the same code-path bug or a single mergeable PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 23, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45548-2", + "nodes": [ + "pull_request:45548", + "pull_request:45662" + ], + "soft_pairs": [ + "pull_request:45548|pull_request:45662" + ], + "prepared_review_unit_hash": "e4a3aa5d077dbee574cdb56d9e83214e86ef7898aa3131afdef3112b68db9304", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4507, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4507, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are about expert-parallel model loading, but they fix different failure modes: one removes `distributed_config` from serialization and adds `has_ep`, while the other prevents parameters from being overwritten during distributed/FSDP2 loading. They are related, not duplicates.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45662 is the more concrete/loading-path-specific fix, with changes in the distributed tensor and module loading flow to prevent rank-0 broadcast from overwriting experts.", + "best_issue_reason": null, + "best_pr_reason": "PR #45662 best represents the cluster because it targets a specific, reproducible bug in EP + FSDP2 loading rather than a broader configuration/serialization adjustment.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "They share the EP/loading theme, but the bugs differ: #45548 fixes config serialization and an EP detection helper, while #45662 fixes parameter clobbering during distributed shard loading. Not the same concrete change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 24, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45639-2", + "nodes": [ + "pull_request:45639", + "pull_request:45645" + ], + "soft_pairs": [ + "pull_request:45639|pull_request:45645" + ], + "prepared_review_unit_hash": "c318cbf72ebd464069a6e844acbc7f768a1830d544eeae979b525ffab97761a0", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4672, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4672, + "estimated_input_tokens": 1168, + "estimated_eval_tokens": 2592 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same xdist-related captured_info collision problem in testing_utils; PR 45645 appears to be the broader end-to-end version, adding the CI-side aggregation needed to consume multiple worker-specific logs.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45645 is the better canonical representative because it includes the underlying xdist-safe logging fix plus the CI workflow update to collect multiple captured_info files, making it the more complete solution.", + "best_issue_reason": null, + "best_pr_reason": "PR 45645 is the strongest overall PR candidate: it fixes the same captured_info collision bug and also preserves CI debug-log visibility when logs are split across workers.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Both PRs fix the same concrete xdist collision for captured_info/debug logs in testing_utils; 45645 is a superset that also updates CI log collection, so they can plausibly be merged into one change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45720-2", + "nodes": [ + "pull_request:45720", + "pull_request:46259" + ], + "soft_pairs": [ + "pull_request:45720|pull_request:46259" + ], + "prepared_review_unit_hash": "a2d04d6554b3eaeefd6495a179783e45bfd20ca31d58a605fc32a6d44955047e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7232, + "estimated_input_tokens": 1808, + "estimated_eval_tokens": 3872 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7232, + "estimated_input_tokens": 1808, + "estimated_eval_tokens": 3872 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two linter-related PRs, but they address different changes: TRF003 adjusts forward signatures/tuple return handling, while TRF018 changes model weight initialization across many architectures.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45720 is the cleaner representative if one must be chosen: it is narrower and targets a single concrete forward-API refactor across model files.", + "best_issue_reason": null, + "best_pr_reason": "PR 45720 is the better standalone representative because it is more focused and self-contained than the broader initialization-rule PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "Different concrete changes and different linter rules: TRF003 adds can_return_tuple/remove return_dict handling, while TRF018 rewrites initialization logic. Shared model filenames are incidental, not the same underlying fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 26, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45737-2", + "nodes": [ + "pull_request:45737", + "pull_request:45738" + ], + "soft_pairs": [ + "pull_request:45737|pull_request:45738" + ], + "prepared_review_unit_hash": "cc2ade73a49ec6a8883a7a24c433988645588bcd049e8aad2bce56b1b24daa1c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same MusicgenMelody cache initialization bug in `forward()`: replacing `EncoderDecoderCache(DynamicCache, DynamicCache)` with a single `DynamicCache`. PR #45737 is the more complete duplicate because it bundles the fix with a regression test, while #45738 is the fix-only variant.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45737 is the best canonical PR because it contains the same code fix as #45738 plus a regression test, making it the most complete representation of the bug fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45737 is the strongest overall candidate: it fixes the concrete cache bug and adds a targeted regression test, whereas #45738 only changes the implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45737", + "right": "pull_request:45738", + "accept": true, + "reason": "Same underlying code-path bug in `musicgen_melody` cache setup; the diffs match on the core fix and the test-only addition in #45737 could plausibly live in the same PR as #45738." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: both PRs make the same `musicgen_melody` cache initialization change, and #45737 only adds a regression test on top of that fix. The soft-edge accept is conservative enough because the diffs are mergeable into one PR for the same bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45932-2", + "nodes": [ + "pull_request:45932", + "pull_request:46100" + ], + "soft_pairs": [ + "pull_request:45932|pull_request:46100" + ], + "prepared_review_unit_hash": "b2184499c222b3a0ab3f5ecdf09e302fbcfd9f99d93f25744497d95dfa605967", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5160, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5160, + "estimated_input_tokens": 1290, + "estimated_eval_tokens": 2836 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two pull requests are not duplicates: one is a CI/telemetry and model-mapping change, while the other is a test migration adding kernel tests and related fixtures. They do not appear to fix the same concrete code-path problem or belong in one merged PR.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "No single PR is a good canonical representative because the PRs are unrelated in scope and purpose: PR 45932 is about CI/secrets and model fixes, while PR 46100 is about test migration and new kernel test coverage.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Different underlying changes and code paths; one is CI/model maintenance, the other is test migration and new kernel tests. They should not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 28, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46126-2", + "nodes": [ + "pull_request:46126", + "pull_request:46269" + ], + "soft_pairs": [ + "pull_request:46126|pull_request:46269" + ], + "prepared_review_unit_hash": "e1c9a3274e67a1fb1e4848b1f941058c3408f4b703c867175cb98c207f0ace0b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5526, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5526, + "estimated_input_tokens": 1382, + "estimated_eval_tokens": 3020 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs overlap on distributed-model code, but they are not the same change: one is a focused EP + Trainer integration PR, while the other is a broad distributed branch base spanning FSDP, CI, model loading, and multiple integration areas.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46126 is the better canonical representative because it is a focused, self-contained feature PR centered on EP + Trainer integration; #46269 is too broad and appears to bundle many unrelated distributed changes.", + "best_issue_reason": null, + "best_pr_reason": "PR #46126 is the strongest standalone PR in this cluster: it has a clear single purpose and coherent scope, whereas #46269 is a large umbrella branch-base change that is not suitable as a duplicate target.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "They share some distributed filenames, but the scopes differ materially. #46126 is a specific EP + Trainer integration change; #46269 is a broad distributed branch base touching FSDP, CI, model loading, and other subsystems. They do not look like the same concrete bug or change and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 29, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46138-10", + "nodes": [ + "pull_request:46138", + "pull_request:46432", + "pull_request:46501", + "pull_request:46502", + "pull_request:46503", + "pull_request:46504", + "pull_request:46554", + "pull_request:46558", + "pull_request:46563", + "pull_request:46716" + ], + "soft_pairs": [ + "pull_request:46138|pull_request:46554", + "pull_request:46138|pull_request:46558", + "pull_request:46138|pull_request:46563", + "pull_request:46138|pull_request:46716", + "pull_request:46138|pull_request:46504", + "pull_request:46432|pull_request:46503", + "pull_request:46432|pull_request:46501", + "pull_request:46432|pull_request:46502", + "pull_request:46432|pull_request:46554", + "pull_request:46432|pull_request:46558", + "pull_request:46501|pull_request:46503", + "pull_request:46501|pull_request:46502", + "pull_request:46501|pull_request:46504", + "pull_request:46501|pull_request:46554", + "pull_request:46502|pull_request:46503", + "pull_request:46502|pull_request:46504", + "pull_request:46502|pull_request:46554", + "pull_request:46503|pull_request:46504", + "pull_request:46503|pull_request:46554", + "pull_request:46504|pull_request:46554", + "pull_request:46504|pull_request:46558", + "pull_request:46554|pull_request:46558", + "pull_request:46554|pull_request:46563", + "pull_request:46554|pull_request:46716", + "pull_request:46558|pull_request:46563", + "pull_request:46558|pull_request:46716", + "pull_request:46563|pull_request:46716", + "pull_request:46504|pull_request:46563", + "pull_request:46504|pull_request:46716" + ], + "prepared_review_unit_hash": "88a1d1cd228e06bc1cf1be3ecc4fb0fb82711d89edd1fde944ab8f07377c4f0f", + "component_budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 29, + "serialized_chars": 25932, + "estimated_input_tokens": 6483, + "estimated_eval_tokens": 13222 + }, + "budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 29, + "serialized_chars": 25932, + "estimated_input_tokens": 6483, + "estimated_eval_tokens": 13222 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Mostly unrelated workflow chore PRs. The only tight cluster is pr-ci-security-gate.yml, but those PRs change different security-gate behaviors, so none are true duplicates.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 46554 is the most substantive representative because it changes the actual security-gate logic, while the others are smaller permission/checkout/secrets plumbing tweaks in different workflows.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46138", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different problems: stdin handling for self-comment CI vs security-gate logic." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46558", + "accept": false, + "reason": "Unrelated workflow changes; one updates comment parsing, the other adjusts PR CI gate permissions/approval flow." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46563", + "accept": false, + "reason": "Different code paths and purposes; not the same underlying fix." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46716", + "accept": false, + "reason": "No shared bug/change beyond both being workflow edits." + }, + { + "left": "pull_request:46138", + "right": "pull_request:46504", + "accept": false, + "reason": "Self-comment CI message piping vs build CI docker tag detection; unrelated." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46503", + "accept": false, + "reason": "SSH runner hardening is unrelated to benchmark dispatch support." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46501", + "accept": false, + "reason": "Different workflows and different concerns; no duplicate bug/change." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46502", + "accept": false, + "reason": "SSH runner changes do not match reviewer assignment workflow changes." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflows and different fixes; not mergeable as one change." + }, + { + "left": "pull_request:46432", + "right": "pull_request:46558", + "accept": false, + "reason": "SSH runner hardening vs PR CI gate permissions/approval logic are unrelated." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46503", + "accept": false, + "reason": "Both are workflow chores, but they change different files and different behavior." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46502", + "accept": false, + "reason": "Anti-slop permission scoping and reviewer assignment ref checkout are different changes." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46504", + "accept": false, + "reason": "Anti-slop workflow vs build CI docker image trigger logic; unrelated." + }, + { + "left": "pull_request:46501", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow files and different security/automation behavior." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46503", + "accept": false, + "reason": "Different workflows and unrelated behavior changes." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46504", + "accept": false, + "reason": "Reviewer assignment ref checkout is unrelated to build image trigger parsing." + }, + { + "left": "pull_request:46502", + "right": "pull_request:46554", + "accept": false, + "reason": "No shared underlying bug; reviewer workflow vs security-gate workflow." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46504", + "accept": false, + "reason": "Benchmark dispatch support and build-image trigger parsing are different concerns." + }, + { + "left": "pull_request:46503", + "right": "pull_request:46554", + "accept": false, + "reason": "Different workflow and different code-path problem." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46554", + "accept": false, + "reason": "Build-ci-docker-images trigger parsing vs PR security gate logic are unrelated." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46558", + "accept": false, + "reason": "Different workflow files and different behavior." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46558", + "accept": false, + "reason": "Same file, but different fixes: one adds diff-based security checks; the other changes permissions and approval gating." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46563", + "accept": false, + "reason": "Same workflow family, but different underlying changes: local checks vs reusable workflow migration." + }, + { + "left": "pull_request:46554", + "right": "pull_request:46716", + "accept": false, + "reason": "Different security hardening steps: local checks vs secrets scoping." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46563", + "accept": false, + "reason": "Different PR CI gate modifications with different code paths and intent." + }, + { + "left": "pull_request:46558", + "right": "pull_request:46716", + "accept": false, + "reason": "Permissions/approval fix is not the same as changing reusable workflow secrets handling." + }, + { + "left": "pull_request:46563", + "right": "pull_request:46716", + "accept": false, + "reason": "Both touch the same workflow file, but one changes workflow source and the other removes inherited secrets; distinct fixes." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46563", + "accept": false, + "reason": "Build image trigger parsing vs security-gate workflow migration are unrelated." + }, + { + "left": "pull_request:46504", + "right": "pull_request:46716", + "accept": false, + "reason": "No common underlying change beyond being workflow edits." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 30, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46174-2", + "nodes": [ + "pull_request:46174", + "pull_request:46780" + ], + "soft_pairs": [ + "pull_request:46174|pull_request:46780" + ], + "prepared_review_unit_hash": "a48ce4ca162f555682f79411d7229ac1ccc08d7822dfab654bb881639fccddbc", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6030, + "estimated_input_tokens": 1508, + "estimated_eval_tokens": 3272 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6030, + "estimated_input_tokens": 1508, + "estimated_eval_tokens": 3272 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs target the same reported MPS crash in the sinusoidal position embedding helper, but they implement different fixes and are not safe to treat as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46174 is the broader, more complete fix for the float64-on-MPS crash: it updates the helper across all affected model files and keeps the internal math on CPU float64 before casting back.", + "best_issue_reason": null, + "best_pr_reason": "PR 46174 is the best representative of the cluster because it is the original, more comprehensive fix and has more discussion/review context than the narrower follow-up.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46174", + "right": "pull_request:46780", + "accept": false, + "reason": "They share the same issue target and helper, but the patches are different solutions: 46174 avoids float64 on MPS by doing the arithmetic on CPU first, while 46780 switches the helper to use dtype for internal arithmetic. Those are alternative implementations, not one mergeable duplicate PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 31, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46310-2", + "nodes": [ + "pull_request:46310", + "pull_request:46546" + ], + "soft_pairs": [ + "pull_request:46310|pull_request:46546" + ], + "prepared_review_unit_hash": "d17fa1e3a8dc4af4c01a1475171e08a83da7a28e6662ffec301a9f7ec75bc370", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4850, + "estimated_input_tokens": 1213, + "estimated_eval_tokens": 2682 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4850, + "estimated_input_tokens": 1213, + "estimated_eval_tokens": 2682 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs fix the same sliding-window right-boundary off-by-one bug in Moonshine streaming; #46546 is the broader, more complete patch and also includes the related Gemma4 fix and tests.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46546 is the better canonical PR because it contains the Moonshine streaming boundary fix plus the matching Gemma4 update and test coverage, making it the most complete representative of this bug-fix cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR #46546 is the best representative overall because it addresses the same concrete boundary bug as #46310, but with broader coverage and validation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46310", + "right": "pull_request:46546", + "accept": true, + "reason": "Same underlying off-by-one in the right-window mask; #46546 is effectively the superset patch with the same Moonshine fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 32, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46376-5", + "nodes": [ + "pull_request:46376", + "pull_request:46417", + "pull_request:46425", + "pull_request:46509", + "pull_request:46917" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417", + "pull_request:46376|pull_request:46425", + "pull_request:46425|pull_request:46509", + "pull_request:46425|pull_request:46917" + ], + "prepared_review_unit_hash": "4ec1bb05fc23428202bfd951785ab2a3a3f8f251fc70752a38e45827ceefb853", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 4, + "serialized_chars": 13785, + "estimated_input_tokens": 3447, + "estimated_eval_tokens": 7150 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 4, + "serialized_chars": 13785, + "estimated_input_tokens": 3447, + "estimated_eval_tokens": 7150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is heterogeneous: one continuous-batching multimodal feature PR, three separate model-integration PRs, and one deprecation cleanup PR. They do not look like duplicate reports of the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46425 is the best single representative if one must be chosen: it is a self-contained model integration with docs, auto-mapping updates, and tests, and it is not a draft.", + "best_issue_reason": null, + "best_pr_reason": "PR 46425 is the most complete and conventional addition in the set; the others are either draft, narrower, or unrelated cleanup.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Different changes and codepaths: multimodal continuous-batching infrastructure vs adding the HunYuan VL model package, config, processor, and docs." + }, + { + "left": "pull_request:46376", + "right": "pull_request:46425", + "accept": false, + "reason": "Different features: continuous-batching multimodal support vs RADIO model integration; they are not the same bug or change." + }, + { + "left": "pull_request:46425", + "right": "pull_request:46509", + "accept": false, + "reason": "Both are model-related but for different architectures and implementations (RADIO vs Nemotron Omni). They would not plausibly be merged into one PR as a single fix/change." + }, + { + "left": "pull_request:46425", + "right": "pull_request:46917", + "accept": false, + "reason": "Unrelated scope: new RADIO model support versus removing deprecated training/image-processing args. Similar repository area is not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 33, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46574-2", + "nodes": [ + "pull_request:46574", + "pull_request:46580" + ], + "soft_pairs": [ + "pull_request:46574|pull_request:46580" + ], + "prepared_review_unit_hash": "093f867ad3150fc37af1691781aebc37414ada8597f47434f874a1910c5ee219", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4409, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4409, + "estimated_input_tokens": 1103, + "estimated_eval_tokens": 2462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are both tokenizer-related but address different bugs in different code paths: one fixes CodeLlama decode stripping, the other fixes DeepSeek checkpoint-to-tokenizer-class resolution. They are not the same underlying change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:46574", + "right": "pull_request:46580", + "accept": false, + "reason": "Different models, different failure modes, and different fixes: CodeLlama decode-space stripping vs DeepSeek auto-tokenizer class misloading. They would not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 34, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46603-2", + "nodes": [ + "pull_request:46603", + "pull_request:46604" + ], + "soft_pairs": [ + "pull_request:46603|pull_request:46604" + ], + "prepared_review_unit_hash": "c2e4cff7505a676873262bb712444840534433031ef73908b3bdac6c73ee3ed5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5496, + "estimated_input_tokens": 1374, + "estimated_eval_tokens": 3004 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5496, + "estimated_input_tokens": 1374, + "estimated_eval_tokens": 3004 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are Mistral tokenizer-related, but they target different changes: one moves the existing MistralConverter into the integrations package, while the other adds tekken tokenizer detection/conversion/save utilities. They share files because they touch the same subsystem, not because they fix the same underlying problem.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: 46603 is a refactor/package move, while 46604 introduces separate tekken tokenizer functionality.", + "best_issue_reason": null, + "best_pr_reason": "46604 is the more substantial and self-contained change, so it is the better representative PR for this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46603", + "right": "pull_request:46604", + "accept": false, + "reason": "Reject: they are adjacent Mistral tokenizer changes, but 46603 only relocates MistralConverter whereas 46604 adds tekken tokenizer detection/conversion/save utilities. Different code-path problem, not mergeable as one duplicate PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 35, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46907-9", + "nodes": [ + "pull_request:46907", + "pull_request:46908", + "pull_request:46926", + "pull_request:46927", + "pull_request:46928", + "pull_request:46941", + "pull_request:46949", + "pull_request:46950", + "pull_request:46951" + ], + "soft_pairs": [ + "pull_request:46907|pull_request:46926", + "pull_request:46907|pull_request:46941", + "pull_request:46907|pull_request:46908", + "pull_request:46907|pull_request:46927", + "pull_request:46907|pull_request:46928", + "pull_request:46908|pull_request:46926", + "pull_request:46908|pull_request:46928", + "pull_request:46908|pull_request:46951", + "pull_request:46908|pull_request:46941", + "pull_request:46926|pull_request:46941", + "pull_request:46926|pull_request:46928", + "pull_request:46926|pull_request:46927", + "pull_request:46927|pull_request:46949", + "pull_request:46927|pull_request:46951", + "pull_request:46927|pull_request:46950", + "pull_request:46927|pull_request:46941", + "pull_request:46928|pull_request:46941", + "pull_request:46928|pull_request:46951", + "pull_request:46949|pull_request:46951", + "pull_request:46949|pull_request:46950", + "pull_request:46907|pull_request:46949", + "pull_request:46908|pull_request:46949", + "pull_request:46950|pull_request:46951", + "pull_request:46908|pull_request:46950", + "pull_request:46926|pull_request:46950" + ], + "prepared_review_unit_hash": "94716d05c63fadf07139b5456dc1eea68ff91ae51ad953e4525fee0a8e3bb7eb", + "component_budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 25, + "serialized_chars": 24934, + "estimated_input_tokens": 6234, + "estimated_eval_tokens": 12724 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 25, + "serialized_chars": 24934, + "estimated_input_tokens": 6234, + "estimated_eval_tokens": 12724 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is heterogeneous: each PR targets a different model or subsystem, and the apparent overlaps are mostly superficial test-fix patterns rather than the same underlying bug or change.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true duplicate PR stands out; if a representative is needed, 46949 is the most central by similarity and impact, but the set is too mixed to form a single canonical change.", + "best_issue_reason": null, + "best_pr_reason": "46941 is the cleanest standalone fix: it corrects a specific DAC distance-calculation bug in model code with a direct output-mismatch symptom.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46907", + "right": "pull_request:46926", + "accept": false, + "reason": "Different models and different failure contexts; Edgetam kwarg cleanup is unrelated to MiniMax test load workarounds." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46941", + "accept": false, + "reason": "Edgetam processor-kwarg handling and DAC distance math fix are unrelated code paths." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46908", + "accept": false, + "reason": "Unrelated fixes: Edgetam model kwargs vs DeepSeek test offload folder for loading." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46927", + "accept": false, + "reason": "Different models and bugs; no shared concrete code-path problem." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46928", + "accept": false, + "reason": "Edgetam runtime kwarg cleanup is unrelated to BLIP-2 tokenizer registration." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46926", + "accept": false, + "reason": "Both use offload_folder in tests, but they are separate model suites and not the same underlying bug/change." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46928", + "accept": false, + "reason": "DeepSeek loading workaround and BLIP-2 tokenizer mapping fix are unrelated." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46951", + "accept": false, + "reason": "Different subsystems and failure modes; test offload workaround is not PEFT hotswap loading logic." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46941", + "accept": false, + "reason": "Separate model test workaround vs DAC model arithmetic fix." + }, + { + "left": "pull_request:46926", + "right": "pull_request:46941", + "accept": false, + "reason": "MiniMax test loading workaround and DAC output-mismatch bug are unrelated." + }, + { + "left": "pull_request:46926", + "right": "pull_request:46928", + "accept": false, + "reason": "Different model families and different fixes; no common code-path." + }, + { + "left": "pull_request:46926", + "right": "pull_request:46927", + "accept": false, + "reason": "MiniMax offload-folder test fix is unrelated to GLM46V input-id handling." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46949", + "accept": false, + "reason": "GLM46V input-id casting and audio/whisper cache-state initialization are different bugs." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46951", + "accept": false, + "reason": "Unrelated model-input casting fix vs PEFT hotswap mismatch handling." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46950", + "accept": false, + "reason": "GLM46V dtype/input handling is unrelated to Mamba2 OOM/kernel behavior." + }, + { + "left": "pull_request:46927", + "right": "pull_request:46941", + "accept": false, + "reason": "Different models and different concrete bugs; not mergeable as one PR." + }, + { + "left": "pull_request:46928", + "right": "pull_request:46941", + "accept": false, + "reason": "Tokenizer auto-mapping change is unrelated to DAC distance computation." + }, + { + "left": "pull_request:46928", + "right": "pull_request:46951", + "accept": false, + "reason": "BLIP-2 tokenizer pattern addition and PEFT loading fix are unrelated." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46951", + "accept": false, + "reason": "Audio/Whisper cache-state initialization and PEFT hotswap loading are different code paths." + }, + { + "left": "pull_request:46949", + "right": "pull_request:46950", + "accept": false, + "reason": "Different model fixes: cache-state initialization vs Mamba2 execution/oom behavior." + }, + { + "left": "pull_request:46907", + "right": "pull_request:46949", + "accept": false, + "reason": "No shared underlying bug; Edgetam kwarg filtering is unrelated to audio/Whisper cache handling." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46949", + "accept": false, + "reason": "Different model test workaround vs audio/Whisper runtime fix." + }, + { + "left": "pull_request:46950", + "right": "pull_request:46951", + "accept": false, + "reason": "Mamba2 kernel/oom fix and PEFT mismatched-size loading fix are unrelated." + }, + { + "left": "pull_request:46908", + "right": "pull_request:46950", + "accept": false, + "reason": "DeepSeek test load workaround and Mamba2 model/runtime change are unrelated." + }, + { + "left": "pull_request:46926", + "right": "pull_request:46950", + "accept": false, + "reason": "MiniMax load test workaround and Mamba2 execution fix are different problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 36, + "component_count": 36, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46933-2", + "nodes": [ + "pull_request:46933", + "pull_request:46942" + ], + "soft_pairs": [ + "pull_request:46933|pull_request:46942" + ], + "prepared_review_unit_hash": "cc88422124b3168df356e35d4c57a6b7ccf179591983300e1759016d35695390", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4371, + "estimated_input_tokens": 1093, + "estimated_eval_tokens": 2442 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4371, + "estimated_input_tokens": 1093, + "estimated_eval_tokens": 2442 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs implement the same coding pattern fix for quantized weights, but they target different model families and different code paths, so they are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46933 is the better canonical representative because it has an explicit linked issue and a clearly scoped fix for Gemma4\u2019s quantized vision/audio embedders.", + "best_issue_reason": null, + "best_pr_reason": "PR #46933 is the strongest standalone representative of the cluster: it is explicit, narrowly scoped, and clearly describes the dtype-cast bug it fixes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46933", + "right": "pull_request:46942", + "accept": false, + "reason": "They share a general quantization dtype-cast pattern, but the concrete bugs are in different models and code paths (Gemma4 embedders vs TimesFM 2.5 ResidualBlock), with different filenames and no evidence they could be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/manifest.json b/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..c7b6480973e6bb171b2c8cb90549d30341078bac --- /dev/null +++ b/snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260630t000016z", + "artifacts": { + "hybrid": "snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-06-30T00:09:42Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260630T000016Z", + "variant": "hybrid" +} diff --git a/snapshots/20260630T000016Z/manifest.json b/snapshots/20260630T000016Z/manifest.json index 2b2a020addd8e04257e22bad6c8ceb91b51aa690..5bac4be3ae4b3f048424803b1c000be51a2093c2 100644 --- a/snapshots/20260630T000016Z/manifest.json +++ b/snapshots/20260630T000016Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 620 }, "extracted_at": "2026-06-30T00:00:16Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260630t000016z", + "runs": { + "hybrid-model-20260630t000016z": { + "analysis_id": "hybrid-model-20260630t000016z", + "artifacts": { + "hybrid": "snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260630T000016Z/analysis-runs/hybrid-model-20260630t000016z/manifest.json", + "model": null, + "published_at": "2026-06-30T00:09:42Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260630T000016Z", "watermark": {